Overview

If I could predict the stock market with even as much as 30% error, I'd be rich enough to inherit the entire nation of Wakanda, vibranium mountain and all.

There's a reason why an entire district of New York City is dedicated to the art of taming the stock market. Whoever cracks the code even slightly has the potential to get absurd levels of returns, and with those returns, lots of bills.

I'll admit the title of this article is a bit misleading. There is no current method to even remotely accurately predict the stock market, but that's kind of the point.

Talk about market efficiency and the paradox of passive investors and active investors

Predicting the market with Stocker
Predicting the market with linear regression
Predicting the market with neural networks
Conclusion -> it's safer to invest in ETFs

The Stocker Module

!git clone https://github.com/WillKoehrsen/Data-Analysis.git

import os

!ls

Data-Analysis  gdrive  sample_data

os.chdir('Data-Analysis')

!ls

 additive_models	    learning_skills	      sentdex_data_analysis
'bayesian drake equation'   LICENSE		      setup
 bayesian_log_reg	    logistic_regression       slack_interaction
 bayesian_lr		    medium		      sp500tickers.pickle
 copernican		    nyc_traffic_data	      statistical_significance
 cyclical-features	    over_vs_under	      statistics
 data-science-tips	    pairplots		      stocker
 datashader-work	    plotly		      think_complexity
 distributions		    poisson		      time_features
 ecdf			    prediction-intervals      univariate_dist
 economics		    random_forest_explained   web_automation
 example_notebook	    README.md		      weighter
 Facts			    recall_precision	      weight_loss_challenge
 geo			    requirements.txt	      widgets

os.chdir('stocker')

!ls

 data	  __pycache__			 'Stocker Prediction Usage.ipynb'
 dev	  readme.md			  stocker.py
 images  'Stocker Analysis Usage.ipynb'

!pip install quandl
!pip install pytrends

import stocker

from stocker import Stocker

goog = Stocker('GOOGL')

goog.plot_stock()

Maximum Adj. Close = 1187.56 on 2018-01-26 00:00:00.
Minimum Adj. Close = 50.16 on 2004-09-03 00:00:00.
Current Adj. Close = 1006.94 on 2018-03-27 00:00:00.

If you pay attention, you'll notice that the dates for the Stocker object are not up-to-date. It stops at 2018--3-27. Taking close look at the actual module code, we'll see that the data is taken from Quandl's WIKI exchange. Perhaps the data is not kept up to date?

We can use Stocker to conduct technical stock analysis, but for now we will focus on being mediums. Stocker uses a package created by Facebook called prophet which is good for additive modeling.

model, model_data = goog.create_prophet_model(days=90)
# Let's make some predictions

Predicted Price on 2018-06-25 00:00:00 = $1175.64

Evaluation of Predictions

Now we will test the stocker predictions. We need to create a test set and a training set. We'll have our training set to be 2014-2016, and our test set to be 2017. Let's see how accurate this model is.

goog.evaluate_prediction()

Prediction Range: 2017-03-27 00:00:00 to 2018-03-27 00:00:00.

Predicted price on 2018-03-24 00:00:00 = $1037.84.
Actual price on    2018-03-23 00:00:00 = $1026.55.

Average Absolute Error on Training Data = $14.31.
Average Absolute Error on Testing  Data = $65.52.

When the model predicted an increase, the price increased 58.27% of the time.
When the model predicted a  decrease, the price decreased  48.18% of the time.

The actual value was within the 80% confidence interval 63.60% of the time.

This is absolutely horrible!

We'll try again, and this time we'll adjust some hyperparameters.

Adjusting the Changepoint Priors

# changepoint priors is the list of changepoints to evaluate

goog.changepoint_prior_analysis(changepoint_priors=[0.001, 0.05, 0.1, 0.2])

goog.changepoint_prior_validation(start_date='2016-01-04', end_date='2017-01-03', changepoint_priors=[0.001, 0.05, 0.1, 0.2])

Validation Range 2016-01-04 00:00:00 to 2017-01-03 00:00:00.

     cps  train_err  train_range    test_err  test_range
0  0.001  39.201088   119.332565   39.458360  119.175959
1  0.050  12.057805    39.832086  204.848301  183.146491
2  0.100  10.765865    36.778261  222.082001  285.733082
3  0.200   9.441183    33.703218  211.435183  491.831858

Evaluating Refined Model

goog.evaluate_prediction()

Prediction Range: 2017-03-27 00:00:00 to 2018-03-27 00:00:00.

Predicted price on 2018-03-24 00:00:00 = $1025.88.
Actual price on    2018-03-23 00:00:00 = $1026.55.

Average Absolute Error on Training Data = $10.92.
Average Absolute Error on Testing  Data = $72.17.

When the model predicted an increase, the price increased 58.74% of the time.
When the model predicted a  decrease, the price decreased  49.06% of the time.

The actual value was within the 80% confidence interval 80.00% of the time.

Testing Our Luck in the Stock Market

Let's see how well our forecasts would play out in the real stock market.

goog.evaluate_prediction(nshares=1000)

/content/Data-Analysis/stocker/stocker.py:613: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy



You played the stock market in GOOGL from 2017-03-27 00:00:00 to 2018-03-27 00:00:00 with 1000 shares.

When the model predicted an increase, the price increased 58.74% of the time.
When the model predicted a  decrease, the price decreased  49.06% of the time.

The total profit using the Prophet model = $87380.00.
The Buy and Hold strategy profit =         $188040.00.

Thanks for playing the stock market!

This shows that it's better to simply invest for the long term.

Preparing Data for More Machine Learning

# Getting the dataframe of the data

goog_data = goog.make_df('2004-08-19', '2018-03-27')
#2004-08-19 00:00:00 to 2018-03-27 00:00:00.

goog_data.head(50)

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Date	Open	High	Low	Close	Volume	Split Ratio	Adj. Open	Adj. High	Adj. Low	Adj. Close	Adj. Volume	ds	y	Daily Change
0	2004-08-19	100.010	104.06	95.96	100.335	44659000.0	1.0	50.159839	52.191109	48.128568	50.322842	44659000.0	2004-08-19	50.322842	0.163003
1	2004-08-20	101.010	109.08	100.50	108.310	22834300.0	1.0	50.661387	54.708881	50.405597	54.322689	22834300.0	2004-08-20	54.322689	3.661302
2	2004-08-23	110.760	113.48	109.05	109.400	18256100.0	1.0	55.551482	56.915693	54.693835	54.869377	18256100.0	2004-08-23	54.869377	-0.682106
3	2004-08-24	111.240	111.60	103.57	104.870	15247300.0	1.0	55.792225	55.972783	51.945350	52.597363	15247300.0	2004-08-24	52.597363	-3.194862
4	2004-08-25	104.760	108.00	103.88	106.000	9188600.0	1.0	52.542193	54.167209	52.100830	53.164113	9188600.0	2004-08-25	53.164113	0.621920
5	2004-08-26	104.950	107.95	104.66	107.910	7094800.0	1.0	52.637487	54.142132	52.492038	54.122070	7094800.0	2004-08-26	54.122070	1.484583
6	2004-08-27	108.100	108.62	105.69	106.150	6211700.0	1.0	54.217364	54.478169	53.008633	53.239345	6211700.0	2004-08-27	53.239345	-0.978019
7	2004-08-30	105.280	105.49	102.01	102.010	5196700.0	1.0	52.802998	52.908323	51.162935	51.162935	5196700.0	2004-08-30	51.162935	-1.640063
8	2004-08-31	102.320	103.71	102.16	102.370	4917800.0	1.0	51.318415	52.015567	51.238167	51.343492	4917800.0	2004-08-31	51.343492	0.025077
9	2004-09-01	102.700	102.97	99.67	100.250	9138200.0	1.0	51.509003	51.644421	49.989312	50.280210	9138200.0	2004-09-01	50.280210	-1.228793
10	2004-09-02	99.090	102.37	98.94	101.510	15118600.0	1.0	49.698414	51.343492	49.623182	50.912161	15118600.0	2004-09-02	50.912161	1.213747
11	2004-09-03	100.950	101.74	99.32	100.010	5152400.0	1.0	50.631294	51.027517	49.813770	50.159839	5152400.0	2004-09-03	50.159839	-0.471455
12	2004-09-07	101.010	102.00	99.61	101.580	5847500.0	1.0	50.661387	51.157920	49.959219	50.947269	5847500.0	2004-09-07	50.947269	0.285882
13	2004-09-08	100.740	103.03	100.50	102.300	4985600.0	1.0	50.525969	51.674514	50.405597	51.308384	4985600.0	2004-09-08	51.308384	0.782415
14	2004-09-09	102.500	102.71	101.00	102.310	4061700.0	1.0	51.408694	51.514019	50.656371	51.313400	4061700.0	2004-09-09	51.313400	-0.095294
15	2004-09-10	101.470	106.56	101.30	105.330	8698800.0	1.0	50.892099	53.444980	50.806836	52.828075	8698800.0	2004-09-10	52.828075	1.935976
16	2004-09-13	106.630	108.41	106.46	107.500	7844100.0	1.0	53.480088	54.372844	53.394825	53.916435	7844100.0	2004-09-13	53.916435	0.436347
17	2004-09-14	107.440	112.00	106.79	111.490	10828900.0	1.0	53.886342	56.173402	53.560336	55.917612	10828900.0	2004-09-14	55.917612	2.031270
18	2004-09-15	110.560	114.23	110.20	112.000	10713000.0	1.0	55.451172	57.291854	55.270615	56.173402	10713000.0	2004-09-15	56.173402	0.722229
19	2004-09-16	112.340	115.80	111.65	113.970	9266300.0	1.0	56.343928	58.079285	55.997860	57.161452	9266300.0	2004-09-16	57.161452	0.817524
20	2004-09-17	114.420	117.49	113.55	117.490	9472500.0	1.0	57.387149	58.926902	56.950802	58.926902	9472500.0	2004-09-17	58.926902	1.539753
21	2004-09-20	116.950	121.60	116.77	119.360	10628700.0	1.0	58.656066	60.988265	58.565787	59.864797	10628700.0	2004-09-20	59.864797	1.208731
22	2004-09-21	120.200	120.42	117.51	117.840	7228700.0	1.0	60.286097	60.396438	58.936933	59.102444	7228700.0	2004-09-21	59.102444	-1.183654
23	2004-09-22	117.450	119.67	116.81	118.380	7581200.0	1.0	58.906840	60.020277	58.585849	59.373280	7581200.0	2004-09-22	59.373280	0.466440
24	2004-09-23	118.840	122.63	117.02	120.820	8535600.0	1.0	59.603992	61.504860	58.691174	60.597057	8535600.0	2004-09-23	60.597057	0.993065
25	2004-09-24	120.970	124.10	119.76	119.830	9123400.0	1.0	60.672290	62.242136	60.065416	60.100525	9123400.0	2004-09-24	60.100525	-0.571765
26	2004-09-27	119.560	120.88	117.80	118.260	7066100.0	1.0	59.965107	60.627150	59.082382	59.313094	7066100.0	2004-09-27	59.313094	-0.652013
27	2004-09-28	121.150	127.40	120.21	126.860	16929000.0	1.0	60.762568	63.897245	60.291113	63.626409	16929000.0	2004-09-28	63.626409	2.863840
28	2004-09-29	126.530	135.02	126.23	131.080	30516400.0	1.0	63.460898	67.719042	63.310433	65.742942	30516400.0	2004-09-29	65.742942	2.282044
29	2004-09-30	129.899	132.30	129.00	129.600	13758000.0	1.0	65.150614	66.354831	64.699722	65.000651	13758000.0	2004-09-30	65.000651	-0.149963
30	2004-10-01	130.800	134.24	128.90	132.580	15124800.0	1.0	65.602509	67.327835	64.649567	66.495265	15124800.0	2004-10-01	66.495265	0.892756
31	2004-10-04	135.275	136.87	134.03	135.060	13022700.0	1.0	67.846937	68.646906	67.222509	67.739104	13022700.0	2004-10-04	67.739104	-0.107833
32	2004-10-05	134.660	138.53	132.24	138.370	14973200.0	1.0	67.538485	69.479476	66.324738	69.399229	14973200.0	2004-10-05	69.399229	1.860744
33	2004-10-06	137.670	138.45	136.00	137.080	13381400.0	1.0	69.048145	69.439353	68.210559	68.752232	13381400.0	2004-10-06	68.752232	-0.295913
34	2004-10-07	136.560	139.88	136.55	138.850	14115000.0	1.0	68.491426	70.156567	68.486411	69.639972	14115000.0	2004-10-07	69.639972	1.148545
35	2004-10-08	138.730	139.68	137.02	137.730	11069500.0	1.0	69.579786	70.056257	68.722139	69.078238	11069500.0	2004-10-08	69.078238	-0.501548
36	2004-10-11	137.010	138.86	133.85	135.260	10472100.0	1.0	68.717123	69.644987	67.132231	67.839414	10472100.0	2004-10-11	67.839414	-0.877709
37	2004-10-12	134.490	137.61	133.40	137.400	11665500.0	1.0	67.453222	69.018052	66.906534	68.912727	11665500.0	2004-10-12	68.912727	1.459505
38	2004-10-13	143.240	143.55	140.08	140.900	19766200.0	1.0	71.841769	71.997249	70.256876	70.668146	19766200.0	2004-10-13	70.668146	-1.173623
39	2004-10-14	141.020	142.38	138.56	142.000	10442100.0	1.0	70.728332	71.410437	69.494523	71.219849	10442100.0	2004-10-14	71.219849	0.491517
40	2004-10-15	144.950	145.50	141.95	144.110	13194700.0	1.0	72.699416	72.975268	71.194771	72.278116	13194700.0	2004-10-15	72.278116	-0.421301
41	2004-10-18	143.120	149.20	141.21	149.160	14036300.0	1.0	71.781583	74.830996	70.823626	74.810934	14036300.0	2004-10-18	74.810934	3.029351
42	2004-10-19	150.500	152.40	147.35	147.940	18109800.0	1.0	75.483009	76.435950	73.903132	74.199045	18109800.0	2004-10-19	74.199045	-1.283963
43	2004-10-20	147.940	148.99	139.60	140.490	22722600.0	1.0	74.199045	74.725671	70.016133	70.462511	22722600.0	2004-10-20	70.462511	-3.736534
44	2004-10-21	144.130	150.13	141.62	149.380	29149800.0	1.0	72.288147	75.297436	71.029261	74.921275	29149800.0	2004-10-21	74.921275	2.633128
45	2004-10-22	170.435	180.17	164.08	172.430	73710000.0	1.0	85.481373	90.363945	82.294034	86.481962	73710000.0	2004-10-22	86.481962	1.000589
46	2004-10-25	176.280	194.43	172.55	187.400	65462800.0	1.0	88.412922	97.516023	86.542147	93.990139	65462800.0	2004-10-25	93.990139	5.577216
47	2004-10-26	186.449	192.64	180.00	181.800	44569500.0	1.0	93.513166	96.618251	90.278682	91.181468	44569500.0	2004-10-26	91.181468	-2.331698
48	2004-10-27	182.509	189.52	181.77	185.970	26686200.0	1.0	91.537066	95.053421	91.166422	93.272925	26686200.0	2004-10-27	93.272925	1.735858
49	2004-10-28	186.630	194.39	185.60	193.300	29663900.0	1.0	93.603946	97.495961	93.087352	96.949273	29663900.0	2004-10-28	96.949273	3.345327

goog_data = goog_data[['Date', 'Open', 'High', 'Low', 'Close', 'Adj. Close', 'Volume']]

goog_data.head(50)

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Date	Open	High	Low	Close	Adj. Close	Volume
0	2004-08-19	100.010	104.06	95.96	100.335	50.322842	44659000.0
1	2004-08-20	101.010	109.08	100.50	108.310	54.322689	22834300.0
2	2004-08-23	110.760	113.48	109.05	109.400	54.869377	18256100.0
3	2004-08-24	111.240	111.60	103.57	104.870	52.597363	15247300.0
4	2004-08-25	104.760	108.00	103.88	106.000	53.164113	9188600.0
5	2004-08-26	104.950	107.95	104.66	107.910	54.122070	7094800.0
6	2004-08-27	108.100	108.62	105.69	106.150	53.239345	6211700.0
7	2004-08-30	105.280	105.49	102.01	102.010	51.162935	5196700.0
8	2004-08-31	102.320	103.71	102.16	102.370	51.343492	4917800.0
9	2004-09-01	102.700	102.97	99.67	100.250	50.280210	9138200.0
10	2004-09-02	99.090	102.37	98.94	101.510	50.912161	15118600.0
11	2004-09-03	100.950	101.74	99.32	100.010	50.159839	5152400.0
12	2004-09-07	101.010	102.00	99.61	101.580	50.947269	5847500.0
13	2004-09-08	100.740	103.03	100.50	102.300	51.308384	4985600.0
14	2004-09-09	102.500	102.71	101.00	102.310	51.313400	4061700.0
15	2004-09-10	101.470	106.56	101.30	105.330	52.828075	8698800.0
16	2004-09-13	106.630	108.41	106.46	107.500	53.916435	7844100.0
17	2004-09-14	107.440	112.00	106.79	111.490	55.917612	10828900.0
18	2004-09-15	110.560	114.23	110.20	112.000	56.173402	10713000.0
19	2004-09-16	112.340	115.80	111.65	113.970	57.161452	9266300.0
20	2004-09-17	114.420	117.49	113.55	117.490	58.926902	9472500.0
21	2004-09-20	116.950	121.60	116.77	119.360	59.864797	10628700.0
22	2004-09-21	120.200	120.42	117.51	117.840	59.102444	7228700.0
23	2004-09-22	117.450	119.67	116.81	118.380	59.373280	7581200.0
24	2004-09-23	118.840	122.63	117.02	120.820	60.597057	8535600.0
25	2004-09-24	120.970	124.10	119.76	119.830	60.100525	9123400.0
26	2004-09-27	119.560	120.88	117.80	118.260	59.313094	7066100.0
27	2004-09-28	121.150	127.40	120.21	126.860	63.626409	16929000.0
28	2004-09-29	126.530	135.02	126.23	131.080	65.742942	30516400.0
29	2004-09-30	129.899	132.30	129.00	129.600	65.000651	13758000.0
30	2004-10-01	130.800	134.24	128.90	132.580	66.495265	15124800.0
31	2004-10-04	135.275	136.87	134.03	135.060	67.739104	13022700.0
32	2004-10-05	134.660	138.53	132.24	138.370	69.399229	14973200.0
33	2004-10-06	137.670	138.45	136.00	137.080	68.752232	13381400.0
34	2004-10-07	136.560	139.88	136.55	138.850	69.639972	14115000.0
35	2004-10-08	138.730	139.68	137.02	137.730	69.078238	11069500.0
36	2004-10-11	137.010	138.86	133.85	135.260	67.839414	10472100.0
37	2004-10-12	134.490	137.61	133.40	137.400	68.912727	11665500.0
38	2004-10-13	143.240	143.55	140.08	140.900	70.668146	19766200.0
39	2004-10-14	141.020	142.38	138.56	142.000	71.219849	10442100.0
40	2004-10-15	144.950	145.50	141.95	144.110	72.278116	13194700.0
41	2004-10-18	143.120	149.20	141.21	149.160	74.810934	14036300.0
42	2004-10-19	150.500	152.40	147.35	147.940	74.199045	18109800.0
43	2004-10-20	147.940	148.99	139.60	140.490	70.462511	22722600.0
44	2004-10-21	144.130	150.13	141.62	149.380	74.921275	29149800.0
45	2004-10-22	170.435	180.17	164.08	172.430	86.481962	73710000.0
46	2004-10-25	176.280	194.43	172.55	187.400	93.990139	65462800.0
47	2004-10-26	186.449	192.64	180.00	181.800	91.181468	44569500.0
48	2004-10-27	182.509	189.52	181.77	185.970	93.272925	26686200.0
49	2004-10-28	186.630	194.39	185.60	193.300	96.949273	29663900.0

Moving Average

import pandas as pd
import numpy as np

import matplotlib.pyplot as plt
%matplotlib inline

import matplotlib.style
import matplotlib as mpl
mpl.style.use('ggplot')

from matplotlib.pylab import rcParams
rcParams['figure.figsize'] = 20, 10

from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler(feature_range=(0, 1))

# Creating copy of goog_data dataframe for moving averages

df = goog_data

df.head()

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Date	Open	High	Low	Close	Adj. Close	Volume
0	2004-08-19	100.01	104.06	95.96	100.335	50.322842	44659000.0
1	2004-08-20	101.01	109.08	100.50	108.310	54.322689	22834300.0
2	2004-08-23	110.76	113.48	109.05	109.400	54.869377	18256100.0
3	2004-08-24	111.24	111.60	103.57	104.870	52.597363	15247300.0
4	2004-08-25	104.76	108.00	103.88	106.000	53.164113	9188600.0

df['Date'] = pd.to_datetime(df.Date, format='%Y-%m-%d')
df.index = df['Date']

df.head(50)

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Date	Open	High	Low	Close	Adj. Close	Volume
Date
2004-08-19	2004-08-19	100.010	104.06	95.96	100.335	50.322842	44659000.0
2004-08-20	2004-08-20	101.010	109.08	100.50	108.310	54.322689	22834300.0
2004-08-23	2004-08-23	110.760	113.48	109.05	109.400	54.869377	18256100.0
2004-08-24	2004-08-24	111.240	111.60	103.57	104.870	52.597363	15247300.0
2004-08-25	2004-08-25	104.760	108.00	103.88	106.000	53.164113	9188600.0
2004-08-26	2004-08-26	104.950	107.95	104.66	107.910	54.122070	7094800.0
2004-08-27	2004-08-27	108.100	108.62	105.69	106.150	53.239345	6211700.0
2004-08-30	2004-08-30	105.280	105.49	102.01	102.010	51.162935	5196700.0
2004-08-31	2004-08-31	102.320	103.71	102.16	102.370	51.343492	4917800.0
2004-09-01	2004-09-01	102.700	102.97	99.67	100.250	50.280210	9138200.0
2004-09-02	2004-09-02	99.090	102.37	98.94	101.510	50.912161	15118600.0
2004-09-03	2004-09-03	100.950	101.74	99.32	100.010	50.159839	5152400.0
2004-09-07	2004-09-07	101.010	102.00	99.61	101.580	50.947269	5847500.0
2004-09-08	2004-09-08	100.740	103.03	100.50	102.300	51.308384	4985600.0
2004-09-09	2004-09-09	102.500	102.71	101.00	102.310	51.313400	4061700.0
2004-09-10	2004-09-10	101.470	106.56	101.30	105.330	52.828075	8698800.0
2004-09-13	2004-09-13	106.630	108.41	106.46	107.500	53.916435	7844100.0
2004-09-14	2004-09-14	107.440	112.00	106.79	111.490	55.917612	10828900.0
2004-09-15	2004-09-15	110.560	114.23	110.20	112.000	56.173402	10713000.0
2004-09-16	2004-09-16	112.340	115.80	111.65	113.970	57.161452	9266300.0
2004-09-17	2004-09-17	114.420	117.49	113.55	117.490	58.926902	9472500.0
2004-09-20	2004-09-20	116.950	121.60	116.77	119.360	59.864797	10628700.0
2004-09-21	2004-09-21	120.200	120.42	117.51	117.840	59.102444	7228700.0
2004-09-22	2004-09-22	117.450	119.67	116.81	118.380	59.373280	7581200.0
2004-09-23	2004-09-23	118.840	122.63	117.02	120.820	60.597057	8535600.0
2004-09-24	2004-09-24	120.970	124.10	119.76	119.830	60.100525	9123400.0
2004-09-27	2004-09-27	119.560	120.88	117.80	118.260	59.313094	7066100.0
2004-09-28	2004-09-28	121.150	127.40	120.21	126.860	63.626409	16929000.0
2004-09-29	2004-09-29	126.530	135.02	126.23	131.080	65.742942	30516400.0
2004-09-30	2004-09-30	129.899	132.30	129.00	129.600	65.000651	13758000.0
2004-10-01	2004-10-01	130.800	134.24	128.90	132.580	66.495265	15124800.0
2004-10-04	2004-10-04	135.275	136.87	134.03	135.060	67.739104	13022700.0
2004-10-05	2004-10-05	134.660	138.53	132.24	138.370	69.399229	14973200.0
2004-10-06	2004-10-06	137.670	138.45	136.00	137.080	68.752232	13381400.0
2004-10-07	2004-10-07	136.560	139.88	136.55	138.850	69.639972	14115000.0
2004-10-08	2004-10-08	138.730	139.68	137.02	137.730	69.078238	11069500.0
2004-10-11	2004-10-11	137.010	138.86	133.85	135.260	67.839414	10472100.0
2004-10-12	2004-10-12	134.490	137.61	133.40	137.400	68.912727	11665500.0
2004-10-13	2004-10-13	143.240	143.55	140.08	140.900	70.668146	19766200.0
2004-10-14	2004-10-14	141.020	142.38	138.56	142.000	71.219849	10442100.0
2004-10-15	2004-10-15	144.950	145.50	141.95	144.110	72.278116	13194700.0
2004-10-18	2004-10-18	143.120	149.20	141.21	149.160	74.810934	14036300.0
2004-10-19	2004-10-19	150.500	152.40	147.35	147.940	74.199045	18109800.0
2004-10-20	2004-10-20	147.940	148.99	139.60	140.490	70.462511	22722600.0
2004-10-21	2004-10-21	144.130	150.13	141.62	149.380	74.921275	29149800.0
2004-10-22	2004-10-22	170.435	180.17	164.08	172.430	86.481962	73710000.0
2004-10-25	2004-10-25	176.280	194.43	172.55	187.400	93.990139	65462800.0
2004-10-26	2004-10-26	186.449	192.64	180.00	181.800	91.181468	44569500.0
2004-10-27	2004-10-27	182.509	189.52	181.77	185.970	93.272925	26686200.0
2004-10-28	2004-10-28	186.630	194.39	185.60	193.300	96.949273	29663900.0

plt.figure(figsize=(16,8))
plt.plot(df['Date'], df['Adj. Close'], label='Close Price history')

[<matplotlib.lines.Line2D at 0x7f6e00e4acc0>]

# Creating dataframe with date and the target variable

data = df.sort_index(ascending=True, axis=0)
new_data = pd.DataFrame(index=range(0, len(df)), columns=['Date', 'Adj. Close'])

for i in range(0, len(data)):
  new_data['Date'][i] = data['Date'][i]
  new_data['Adj. Close'][i] = data['Adj. Close'][i]

# Train-test split

train = new_data[:2600]
test = new_data[2600:]

new_data.shape, train.shape, test.shape

((3424, 2), (2600, 2), (824, 2))

num = test.shape[0]

train['Date'].min(), train['Date'].max(), test['Date'].min(), test['Date'].max()

(Timestamp('2004-08-19 00:00:00'),
 Timestamp('2014-12-15 00:00:00'),
 Timestamp('2014-12-16 00:00:00'),
 Timestamp('2018-03-27 00:00:00'))

# Making predictions

preds = []
for i in range(0, num):
  a = train['Adj. Close'][len(train)-924+i:].sum() + sum(preds)
  b = a/num
  preds.append(b)

len(preds)

# Measure accuracy with rmse (Root Mean Squared Error)

rms=np.sqrt(np.mean(np.power((np.array(test['Adj. Close'])-preds),2)))

print(rms)

264.46002931639754

test['Predictions'] = 0
test['Predictions'] = preds
plt.plot(train['Adj. Close'])
plt.plot(test[['Adj. Close', 'Predictions']])

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy






[<matplotlib.lines.Line2D at 0x7f6dee712940>,
 <matplotlib.lines.Line2D at 0x7f6dee712b70>]

Simple Linear Regression

lr_data = goog_data

lr_data.head(50)

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Date	Open	High	Low	Close	Adj. Close	Volume
Date
2004-08-19	2004-08-19	100.010	104.06	95.96	100.335	50.322842	44659000.0
2004-08-20	2004-08-20	101.010	109.08	100.50	108.310	54.322689	22834300.0
2004-08-23	2004-08-23	110.760	113.48	109.05	109.400	54.869377	18256100.0
2004-08-24	2004-08-24	111.240	111.60	103.57	104.870	52.597363	15247300.0
2004-08-25	2004-08-25	104.760	108.00	103.88	106.000	53.164113	9188600.0
2004-08-26	2004-08-26	104.950	107.95	104.66	107.910	54.122070	7094800.0
2004-08-27	2004-08-27	108.100	108.62	105.69	106.150	53.239345	6211700.0
2004-08-30	2004-08-30	105.280	105.49	102.01	102.010	51.162935	5196700.0
2004-08-31	2004-08-31	102.320	103.71	102.16	102.370	51.343492	4917800.0
2004-09-01	2004-09-01	102.700	102.97	99.67	100.250	50.280210	9138200.0
2004-09-02	2004-09-02	99.090	102.37	98.94	101.510	50.912161	15118600.0
2004-09-03	2004-09-03	100.950	101.74	99.32	100.010	50.159839	5152400.0
2004-09-07	2004-09-07	101.010	102.00	99.61	101.580	50.947269	5847500.0
2004-09-08	2004-09-08	100.740	103.03	100.50	102.300	51.308384	4985600.0
2004-09-09	2004-09-09	102.500	102.71	101.00	102.310	51.313400	4061700.0
2004-09-10	2004-09-10	101.470	106.56	101.30	105.330	52.828075	8698800.0
2004-09-13	2004-09-13	106.630	108.41	106.46	107.500	53.916435	7844100.0
2004-09-14	2004-09-14	107.440	112.00	106.79	111.490	55.917612	10828900.0
2004-09-15	2004-09-15	110.560	114.23	110.20	112.000	56.173402	10713000.0
2004-09-16	2004-09-16	112.340	115.80	111.65	113.970	57.161452	9266300.0
2004-09-17	2004-09-17	114.420	117.49	113.55	117.490	58.926902	9472500.0
2004-09-20	2004-09-20	116.950	121.60	116.77	119.360	59.864797	10628700.0
2004-09-21	2004-09-21	120.200	120.42	117.51	117.840	59.102444	7228700.0
2004-09-22	2004-09-22	117.450	119.67	116.81	118.380	59.373280	7581200.0
2004-09-23	2004-09-23	118.840	122.63	117.02	120.820	60.597057	8535600.0
2004-09-24	2004-09-24	120.970	124.10	119.76	119.830	60.100525	9123400.0
2004-09-27	2004-09-27	119.560	120.88	117.80	118.260	59.313094	7066100.0
2004-09-28	2004-09-28	121.150	127.40	120.21	126.860	63.626409	16929000.0
2004-09-29	2004-09-29	126.530	135.02	126.23	131.080	65.742942	30516400.0
2004-09-30	2004-09-30	129.899	132.30	129.00	129.600	65.000651	13758000.0
2004-10-01	2004-10-01	130.800	134.24	128.90	132.580	66.495265	15124800.0
2004-10-04	2004-10-04	135.275	136.87	134.03	135.060	67.739104	13022700.0
2004-10-05	2004-10-05	134.660	138.53	132.24	138.370	69.399229	14973200.0
2004-10-06	2004-10-06	137.670	138.45	136.00	137.080	68.752232	13381400.0
2004-10-07	2004-10-07	136.560	139.88	136.55	138.850	69.639972	14115000.0
2004-10-08	2004-10-08	138.730	139.68	137.02	137.730	69.078238	11069500.0
2004-10-11	2004-10-11	137.010	138.86	133.85	135.260	67.839414	10472100.0
2004-10-12	2004-10-12	134.490	137.61	133.40	137.400	68.912727	11665500.0
2004-10-13	2004-10-13	143.240	143.55	140.08	140.900	70.668146	19766200.0
2004-10-14	2004-10-14	141.020	142.38	138.56	142.000	71.219849	10442100.0
2004-10-15	2004-10-15	144.950	145.50	141.95	144.110	72.278116	13194700.0
2004-10-18	2004-10-18	143.120	149.20	141.21	149.160	74.810934	14036300.0
2004-10-19	2004-10-19	150.500	152.40	147.35	147.940	74.199045	18109800.0
2004-10-20	2004-10-20	147.940	148.99	139.60	140.490	70.462511	22722600.0
2004-10-21	2004-10-21	144.130	150.13	141.62	149.380	74.921275	29149800.0
2004-10-22	2004-10-22	170.435	180.17	164.08	172.430	86.481962	73710000.0
2004-10-25	2004-10-25	176.280	194.43	172.55	187.400	93.990139	65462800.0
2004-10-26	2004-10-26	186.449	192.64	180.00	181.800	91.181468	44569500.0
2004-10-27	2004-10-27	182.509	189.52	181.77	185.970	93.272925	26686200.0
2004-10-28	2004-10-28	186.630	194.39	185.60	193.300	96.949273	29663900.0

# We'll create a separate dataset so that new features don't mess up the original data.


lr_data['Date'] = pd.to_datetime(lr_data.Date, format='%Y-%m-%d')
lr_data.index = lr_data['Date']

lr_data = lr_data.sort_index(ascending=True, axis=0)

new_data = pd.DataFrame(index=range(0, len(lr_data)), columns=['Date', 'Adj. Close'])
for i in range(0,len(data)):
    new_data['Date'][i] = lr_data['Date'][i]
    new_data['Adj. Close'][i] = lr_data['Adj. Close'][i]

new_data.head(50)

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Date	Adj. Close
0	2004-08-19 00:00:00	50.3228
1	2004-08-20 00:00:00	54.3227
2	2004-08-23 00:00:00	54.8694
3	2004-08-24 00:00:00	52.5974
4	2004-08-25 00:00:00	53.1641
5	2004-08-26 00:00:00	54.1221
6	2004-08-27 00:00:00	53.2393
7	2004-08-30 00:00:00	51.1629
8	2004-08-31 00:00:00	51.3435
9	2004-09-01 00:00:00	50.2802
10	2004-09-02 00:00:00	50.9122
11	2004-09-03 00:00:00	50.1598
12	2004-09-07 00:00:00	50.9473
13	2004-09-08 00:00:00	51.3084
14	2004-09-09 00:00:00	51.3134
15	2004-09-10 00:00:00	52.8281
16	2004-09-13 00:00:00	53.9164
17	2004-09-14 00:00:00	55.9176
18	2004-09-15 00:00:00	56.1734
19	2004-09-16 00:00:00	57.1615
20	2004-09-17 00:00:00	58.9269
21	2004-09-20 00:00:00	59.8648
22	2004-09-21 00:00:00	59.1024
23	2004-09-22 00:00:00	59.3733
24	2004-09-23 00:00:00	60.5971
25	2004-09-24 00:00:00	60.1005
26	2004-09-27 00:00:00	59.3131
27	2004-09-28 00:00:00	63.6264
28	2004-09-29 00:00:00	65.7429
29	2004-09-30 00:00:00	65.0007
30	2004-10-01 00:00:00	66.4953
31	2004-10-04 00:00:00	67.7391
32	2004-10-05 00:00:00	69.3992
33	2004-10-06 00:00:00	68.7522
34	2004-10-07 00:00:00	69.64
35	2004-10-08 00:00:00	69.0782
36	2004-10-11 00:00:00	67.8394
37	2004-10-12 00:00:00	68.9127
38	2004-10-13 00:00:00	70.6681
39	2004-10-14 00:00:00	71.2198
40	2004-10-15 00:00:00	72.2781
41	2004-10-18 00:00:00	74.8109
42	2004-10-19 00:00:00	74.199
43	2004-10-20 00:00:00	70.4625
44	2004-10-21 00:00:00	74.9213
45	2004-10-22 00:00:00	86.482
46	2004-10-25 00:00:00	93.9901
47	2004-10-26 00:00:00	91.1815
48	2004-10-27 00:00:00	93.2729
49	2004-10-28 00:00:00	96.9493

!pip install fastai==0.7.0

from fastai.structured import add_datepart

add_datepart(new_data, 'Date')
new_data.drop('Elapsed', axis=1, inplace=True)

new_data.head(50)

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Adj. Close	Year	Month	Week	Day	Dayofweek	Dayofyear	Is_month_end	Is_month_start	Is_quarter_end	Is_quarter_start	Is_year_end	Is_year_start
0	50.3228	2004	8	34	19	3	232	False	False	False	False	False	False
1	54.3227	2004	8	34	20	4	233	False	False	False	False	False	False
2	54.8694	2004	8	35	23	0	236	False	False	False	False	False	False
3	52.5974	2004	8	35	24	1	237	False	False	False	False	False	False
4	53.1641	2004	8	35	25	2	238	False	False	False	False	False	False
5	54.1221	2004	8	35	26	3	239	False	False	False	False	False	False
6	53.2393	2004	8	35	27	4	240	False	False	False	False	False	False
7	51.1629	2004	8	36	30	0	243	False	False	False	False	False	False
8	51.3435	2004	8	36	31	1	244	True	False	False	False	False	False
9	50.2802	2004	9	36	1	2	245	False	True	False	False	False	False
10	50.9122	2004	9	36	2	3	246	False	False	False	False	False	False
11	50.1598	2004	9	36	3	4	247	False	False	False	False	False	False
12	50.9473	2004	9	37	7	1	251	False	False	False	False	False	False
13	51.3084	2004	9	37	8	2	252	False	False	False	False	False	False
14	51.3134	2004	9	37	9	3	253	False	False	False	False	False	False
15	52.8281	2004	9	37	10	4	254	False	False	False	False	False	False
16	53.9164	2004	9	38	13	0	257	False	False	False	False	False	False
17	55.9176	2004	9	38	14	1	258	False	False	False	False	False	False
18	56.1734	2004	9	38	15	2	259	False	False	False	False	False	False
19	57.1615	2004	9	38	16	3	260	False	False	False	False	False	False
20	58.9269	2004	9	38	17	4	261	False	False	False	False	False	False
21	59.8648	2004	9	39	20	0	264	False	False	False	False	False	False
22	59.1024	2004	9	39	21	1	265	False	False	False	False	False	False
23	59.3733	2004	9	39	22	2	266	False	False	False	False	False	False
24	60.5971	2004	9	39	23	3	267	False	False	False	False	False	False
25	60.1005	2004	9	39	24	4	268	False	False	False	False	False	False
26	59.3131	2004	9	40	27	0	271	False	False	False	False	False	False
27	63.6264	2004	9	40	28	1	272	False	False	False	False	False	False
28	65.7429	2004	9	40	29	2	273	False	False	False	False	False	False
29	65.0007	2004	9	40	30	3	274	True	False	True	False	False	False
30	66.4953	2004	10	40	1	4	275	False	True	False	True	False	False
31	67.7391	2004	10	41	4	0	278	False	False	False	False	False	False
32	69.3992	2004	10	41	5	1	279	False	False	False	False	False	False
33	68.7522	2004	10	41	6	2	280	False	False	False	False	False	False
34	69.64	2004	10	41	7	3	281	False	False	False	False	False	False
35	69.0782	2004	10	41	8	4	282	False	False	False	False	False	False
36	67.8394	2004	10	42	11	0	285	False	False	False	False	False	False
37	68.9127	2004	10	42	12	1	286	False	False	False	False	False	False
38	70.6681	2004	10	42	13	2	287	False	False	False	False	False	False
39	71.2198	2004	10	42	14	3	288	False	False	False	False	False	False
40	72.2781	2004	10	42	15	4	289	False	False	False	False	False	False
41	74.8109	2004	10	43	18	0	292	False	False	False	False	False	False
42	74.199	2004	10	43	19	1	293	False	False	False	False	False	False
43	70.4625	2004	10	43	20	2	294	False	False	False	False	False	False
44	74.9213	2004	10	43	21	3	295	False	False	False	False	False	False
45	86.482	2004	10	43	22	4	296	False	False	False	False	False	False
46	93.9901	2004	10	44	25	0	299	False	False	False	False	False	False
47	91.1815	2004	10	44	26	1	300	False	False	False	False	False	False
48	93.2729	2004	10	44	27	2	301	False	False	False	False	False	False
49	96.9493	2004	10	44	28	3	302	False	False	False	False	False	False

# Train-test split

train = new_data[:2600]
test = new_data[2600:]

x_train = train.drop('Adj. Close', axis=1)
y_train = train['Adj. Close']
x_test = test.drop('Adj. Close', axis=1)
y_test = test['Adj. Close']

x_train

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	Year	Month	Week	Day	Dayofweek	Dayofyear	Is_month_end	Is_month_start	Is_quarter_end	Is_quarter_start	Is_year_end	Is_year_start
0	2004	8	34	19	3	232	False	False	False	False	False	False
1	2004	8	34	20	4	233	False	False	False	False	False	False
2	2004	8	35	23	0	236	False	False	False	False	False	False
3	2004	8	35	24	1	237	False	False	False	False	False	False
4	2004	8	35	25	2	238	False	False	False	False	False	False
5	2004	8	35	26	3	239	False	False	False	False	False	False
6	2004	8	35	27	4	240	False	False	False	False	False	False
7	2004	8	36	30	0	243	False	False	False	False	False	False
8	2004	8	36	31	1	244	True	False	False	False	False	False
9	2004	9	36	1	2	245	False	True	False	False	False	False
10	2004	9	36	2	3	246	False	False	False	False	False	False
11	2004	9	36	3	4	247	False	False	False	False	False	False
12	2004	9	37	7	1	251	False	False	False	False	False	False
13	2004	9	37	8	2	252	False	False	False	False	False	False
14	2004	9	37	9	3	253	False	False	False	False	False	False
15	2004	9	37	10	4	254	False	False	False	False	False	False
16	2004	9	38	13	0	257	False	False	False	False	False	False
17	2004	9	38	14	1	258	False	False	False	False	False	False
18	2004	9	38	15	2	259	False	False	False	False	False	False
19	2004	9	38	16	3	260	False	False	False	False	False	False
20	2004	9	38	17	4	261	False	False	False	False	False	False
21	2004	9	39	20	0	264	False	False	False	False	False	False
22	2004	9	39	21	1	265	False	False	False	False	False	False
23	2004	9	39	22	2	266	False	False	False	False	False	False
24	2004	9	39	23	3	267	False	False	False	False	False	False
25	2004	9	39	24	4	268	False	False	False	False	False	False
26	2004	9	40	27	0	271	False	False	False	False	False	False
27	2004	9	40	28	1	272	False	False	False	False	False	False
28	2004	9	40	29	2	273	False	False	False	False	False	False
29	2004	9	40	30	3	274	True	False	True	False	False	False
...	...	...	...	...	...	...	...	...	...	...	...	...
2570	2014	11	45	3	0	307	False	False	False	False	False	False
2571	2014	11	45	4	1	308	False	False	False	False	False	False
2572	2014	11	45	5	2	309	False	False	False	False	False	False
2573	2014	11	45	6	3	310	False	False	False	False	False	False
2574	2014	11	45	7	4	311	False	False	False	False	False	False
2575	2014	11	46	10	0	314	False	False	False	False	False	False
2576	2014	11	46	11	1	315	False	False	False	False	False	False
2577	2014	11	46	12	2	316	False	False	False	False	False	False
2578	2014	11	46	13	3	317	False	False	False	False	False	False
2579	2014	11	46	14	4	318	False	False	False	False	False	False
2580	2014	11	47	17	0	321	False	False	False	False	False	False
2581	2014	11	47	18	1	322	False	False	False	False	False	False
2582	2014	11	47	19	2	323	False	False	False	False	False	False
2583	2014	11	47	20	3	324	False	False	False	False	False	False
2584	2014	11	47	21	4	325	False	False	False	False	False	False
2585	2014	11	48	24	0	328	False	False	False	False	False	False
2586	2014	11	48	25	1	329	False	False	False	False	False	False
2587	2014	11	48	26	2	330	False	False	False	False	False	False
2588	2014	11	48	28	4	332	False	False	False	False	False	False
2589	2014	12	49	1	0	335	False	True	False	False	False	False
2590	2014	12	49	2	1	336	False	False	False	False	False	False
2591	2014	12	49	3	2	337	False	False	False	False	False	False
2592	2014	12	49	4	3	338	False	False	False	False	False	False
2593	2014	12	49	5	4	339	False	False	False	False	False	False
2594	2014	12	50	8	0	342	False	False	False	False	False	False
2595	2014	12	50	9	1	343	False	False	False	False	False	False
2596	2014	12	50	10	2	344	False	False	False	False	False	False
2597	2014	12	50	11	3	345	False	False	False	False	False	False
2598	2014	12	50	12	4	346	False	False	False	False	False	False
2599	2014	12	51	15	0	349	False	False	False	False	False	False

2600 rows × 12 columns

# Implementing linear regression
from sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(x_train, y_train)

LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)

# Predictions 
preds = model.predict(x_test)
rms = np.sqrt(np.mean(np.power((np.array(y_test)-np.array(preds)),2)))

print(rms)

292.21562094558186

# Plot

test['Predictions'] = 0
test['Predictions'] = preds


plt.plot(train['Adj. Close'])
plt.plot(test[['Adj. Close', 'Predictions']])

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:3: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy






[<matplotlib.lines.Line2D at 0x7f6dee6f80f0>,
 <matplotlib.lines.Line2D at 0x7f6dee6f82e8>]

k-Nearest Neighbours

from sklearn import neighbors
from sklearn.model_selection import GridSearchCV
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler(feature_range=(0, 1))

# scaling the data

x_train_scaled = scaler.fit_transform(x_train)
x_train = pd.DataFrame(x_train_scaled)
x_test_scaled = scaler.fit_transform(x_test)
x_test = pd.DataFrame(x_test_scaled)

# using gridsearch to find the best value of k

params = {'n_neighbors': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]}
knn = neighbors.KNeighborsRegressor()
model = GridSearchCV(knn, params, cv=5)

# fitting the model and predicting
model.fit(x_train, y_train)
preds = model.predict(x_test)

# Results

rms = np.sqrt(np.mean(np.power((np.array(y_test)-np.array(preds)),2)))

print(rms)

589.3454139574452

test['Predictions'] = 0
test['Predictions'] = new_preds


plt.plot(train['Adj. Close'])
plt.plot(test[['Adj. Close', 'Predictions']])

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy






[<matplotlib.lines.Line2D at 0x7f6dd249ab70>,
 <matplotlib.lines.Line2D at 0x7f6dd24368d0>]

Multilayer Perceptron

import tensorflow as tf
from tensorflow.keras import layers

model = tf.keras.models.Sequential()

model.add(tf.keras.layers.Dense(100, activation=tf.nn.relu))

model.add(tf.keras.layers.Dense(100, activation=tf.nn.relu))

model.add(tf.keras.layers.Dense(1, activation=tf.nn.relu))

model.compile(optimizer='adam', loss='mean_squared_error')

X_train = np.array(x_train)
Y_train = np.array(y_train)

X_train

array([[0.     , 0.63636, 0.63462, ..., 0.     , 0.     , 0.     ],
       [0.     , 0.63636, 0.63462, ..., 0.     , 0.     , 0.     ],
       [0.     , 0.63636, 0.65385, ..., 0.     , 0.     , 0.     ],
       ...,
       [1.     , 0.72727, 0.75   , ..., 0.     , 0.     , 0.     ],
       [1.     , 0.72727, 0.75   , ..., 0.     , 0.     , 0.     ],
       [1.     , 0.81818, 0.75   , ..., 1.     , 0.     , 0.     ]])

X_train.shape

(2800, 12)

model.fit(X_train, Y_train, epochs=500)

Epoch 1/500
2800/2800 [==============================] - 1s 297us/sample - loss: 107042.6812
Epoch 2/500
2800/2800 [==============================] - 0s 67us/sample - loss: 41312.2483
Epoch 3/500
2800/2800 [==============================] - 0s 60us/sample - loss: 18302.2589
Epoch 4/500
2800/2800 [==============================] - 0s 66us/sample - loss: 14739.7240
Epoch 5/500
2800/2800 [==============================] - 0s 65us/sample - loss: 11421.5729
Epoch 6/500
2800/2800 [==============================] - 0s 64us/sample - loss: 8606.6912
Epoch 7/500
2800/2800 [==============================] - 0s 65us/sample - loss: 6625.6495
Epoch 8/500
2800/2800 [==============================] - 0s 62us/sample - loss: 5474.2571
Epoch 9/500
2800/2800 [==============================] - 0s 60us/sample - loss: 4892.3979
Epoch 10/500
2800/2800 [==============================] - 0s 59us/sample - loss: 4633.2238
Epoch 11/500
2800/2800 [==============================] - 0s 62us/sample - loss: 4482.4471
Epoch 12/500
2800/2800 [==============================] - 0s 59us/sample - loss: 4380.0899
Epoch 13/500
2800/2800 [==============================] - 0s 55us/sample - loss: 4300.8690
Epoch 14/500
2800/2800 [==============================] - 0s 62us/sample - loss: 4214.3174
Epoch 15/500
2800/2800 [==============================] - 0s 59us/sample - loss: 4159.0237
Epoch 16/500
2800/2800 [==============================] - 0s 64us/sample - loss: 4102.6403
Epoch 17/500
2800/2800 [==============================] - 0s 62us/sample - loss: 4074.3091
Epoch 18/500
2800/2800 [==============================] - 0s 62us/sample - loss: 4040.7758
Epoch 19/500
2800/2800 [==============================] - 0s 56us/sample - loss: 4000.0518
Epoch 20/500
2800/2800 [==============================] - 0s 56us/sample - loss: 4004.0519
Epoch 21/500
2800/2800 [==============================] - 0s 60us/sample - loss: 3929.9223
Epoch 22/500
2800/2800 [==============================] - 0s 56us/sample - loss: 3899.7539
Epoch 23/500
2800/2800 [==============================] - 0s 60us/sample - loss: 3859.3802
Epoch 24/500
2800/2800 [==============================] - 0s 54us/sample - loss: 3802.6503
Epoch 25/500
2800/2800 [==============================] - 0s 57us/sample - loss: 3777.4047
Epoch 26/500
2800/2800 [==============================] - 0s 60us/sample - loss: 3721.2936
Epoch 27/500
2800/2800 [==============================] - 0s 59us/sample - loss: 3658.5611
Epoch 28/500
2800/2800 [==============================] - 0s 57us/sample - loss: 3596.7238
Epoch 29/500
2800/2800 [==============================] - 0s 63us/sample - loss: 3569.7354
Epoch 30/500
2800/2800 [==============================] - 0s 57us/sample - loss: 3508.0627
Epoch 31/500
2800/2800 [==============================] - 0s 58us/sample - loss: 3432.4962
Epoch 32/500
2800/2800 [==============================] - 0s 59us/sample - loss: 3371.8384
Epoch 33/500
2800/2800 [==============================] - 0s 55us/sample - loss: 3327.3058
Epoch 34/500
2800/2800 [==============================] - 0s 59us/sample - loss: 3239.4707
Epoch 35/500
2800/2800 [==============================] - 0s 60us/sample - loss: 3211.6994
Epoch 36/500
2800/2800 [==============================] - 0s 55us/sample - loss: 3118.9492
Epoch 37/500
2800/2800 [==============================] - 0s 60us/sample - loss: 3059.0352
Epoch 38/500
2800/2800 [==============================] - 0s 54us/sample - loss: 2997.7588
Epoch 39/500
2800/2800 [==============================] - 0s 57us/sample - loss: 2929.7851
Epoch 40/500
2800/2800 [==============================] - 0s 62us/sample - loss: 2875.7662
Epoch 41/500
2800/2800 [==============================] - 0s 60us/sample - loss: 2815.8341
Epoch 42/500
2800/2800 [==============================] - 0s 55us/sample - loss: 2755.2486
Epoch 43/500
2800/2800 [==============================] - 0s 62us/sample - loss: 2743.1073
Epoch 44/500
2800/2800 [==============================] - 0s 55us/sample - loss: 2650.5037
Epoch 45/500
2800/2800 [==============================] - 0s 62us/sample - loss: 2611.3546
Epoch 46/500
2800/2800 [==============================] - 0s 59us/sample - loss: 2559.4127
Epoch 47/500
2800/2800 [==============================] - 0s 62us/sample - loss: 2502.4777
Epoch 48/500
2800/2800 [==============================] - 0s 59us/sample - loss: 2473.7575
Epoch 49/500
2800/2800 [==============================] - 0s 60us/sample - loss: 2427.9666
Epoch 50/500
2800/2800 [==============================] - 0s 55us/sample - loss: 2375.8519
Epoch 51/500
2800/2800 [==============================] - 0s 59us/sample - loss: 2335.6260
Epoch 52/500
2800/2800 [==============================] - 0s 60us/sample - loss: 2302.3150
Epoch 53/500
2800/2800 [==============================] - 0s 63us/sample - loss: 2261.2262
Epoch 54/500
2800/2800 [==============================] - 0s 57us/sample - loss: 2276.2005
Epoch 55/500
2800/2800 [==============================] - 0s 61us/sample - loss: 2217.6836
Epoch 56/500
2800/2800 [==============================] - 0s 53us/sample - loss: 2200.5515
Epoch 57/500
2800/2800 [==============================] - 0s 69us/sample - loss: 2166.6168
Epoch 58/500
2800/2800 [==============================] - 0s 62us/sample - loss: 2149.8538
Epoch 59/500
2800/2800 [==============================] - 0s 62us/sample - loss: 2131.2684
Epoch 60/500
2800/2800 [==============================] - 0s 65us/sample - loss: 2117.5617
Epoch 61/500
2800/2800 [==============================] - 0s 59us/sample - loss: 2136.8635
Epoch 62/500
2800/2800 [==============================] - 0s 59us/sample - loss: 2086.9172
Epoch 63/500
2800/2800 [==============================] - 0s 60us/sample - loss: 2074.4112
Epoch 64/500
2800/2800 [==============================] - 0s 63us/sample - loss: 2070.8835
Epoch 65/500
2800/2800 [==============================] - 0s 65us/sample - loss: 2049.8117
Epoch 66/500
2800/2800 [==============================] - 0s 56us/sample - loss: 2059.9661
Epoch 67/500
2800/2800 [==============================] - 0s 64us/sample - loss: 2042.7345
Epoch 68/500
2800/2800 [==============================] - 0s 58us/sample - loss: 2043.8399
Epoch 69/500
2800/2800 [==============================] - 0s 64us/sample - loss: 2043.6758
Epoch 70/500
2800/2800 [==============================] - 0s 65us/sample - loss: 2010.1117
Epoch 71/500
2800/2800 [==============================] - 0s 57us/sample - loss: 2006.6912
Epoch 72/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1998.4531
Epoch 73/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1984.5140
Epoch 74/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1996.5648
Epoch 75/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1962.1523
Epoch 76/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1955.9846
Epoch 77/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1955.2565
Epoch 78/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1947.7191
Epoch 79/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1953.3301
Epoch 80/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1934.9385
Epoch 81/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1924.2109
Epoch 82/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1910.7859
Epoch 83/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1897.3683
Epoch 84/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1907.4632
Epoch 85/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1891.1020
Epoch 86/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1891.9109
Epoch 87/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1864.7346
Epoch 88/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1862.2231
Epoch 89/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1844.3185
Epoch 90/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1834.2257
Epoch 91/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1833.4358
Epoch 92/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1814.5655
Epoch 93/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1791.7842
Epoch 94/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1780.9216
Epoch 95/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1775.3002
Epoch 96/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1775.6334
Epoch 97/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1741.6012
Epoch 98/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1748.9729
Epoch 99/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1729.0204
Epoch 100/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1717.6277
Epoch 101/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1718.6550
Epoch 102/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1688.5626
Epoch 103/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1678.3398
Epoch 104/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1653.9336
Epoch 105/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1639.7355
Epoch 106/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1652.1106
Epoch 107/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1614.5160
Epoch 108/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1596.8088
Epoch 109/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1614.2061
Epoch 110/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1573.5171
Epoch 111/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1572.3198
Epoch 112/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1559.2580
Epoch 113/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1563.9145
Epoch 114/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1538.0331
Epoch 115/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1522.4755
Epoch 116/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1514.9207
Epoch 117/500
2800/2800 [==============================] - 0s 72us/sample - loss: 1505.6984
Epoch 118/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1497.6740
Epoch 119/500
2800/2800 [==============================] - 0s 67us/sample - loss: 1489.2281
Epoch 120/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1472.0496
Epoch 121/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1455.5400
Epoch 122/500
2800/2800 [==============================] - 0s 68us/sample - loss: 1455.3277
Epoch 123/500
2800/2800 [==============================] - 0s 72us/sample - loss: 1448.1577
Epoch 124/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1442.3502
Epoch 125/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1438.5963
Epoch 126/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1420.4090
Epoch 127/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1410.9522
Epoch 128/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1404.9106
Epoch 129/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1394.6398
Epoch 130/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1394.5640
Epoch 131/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1380.8252
Epoch 132/500
2800/2800 [==============================] - 0s 68us/sample - loss: 1394.0857
Epoch 133/500
2800/2800 [==============================] - 0s 70us/sample - loss: 1379.1661
Epoch 134/500
2800/2800 [==============================] - 0s 68us/sample - loss: 1367.0957
Epoch 135/500
2800/2800 [==============================] - 0s 67us/sample - loss: 1369.8024
Epoch 136/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1372.2572
Epoch 137/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1349.9875
Epoch 138/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1328.4315
Epoch 139/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1345.7265
Epoch 140/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1333.4088
Epoch 141/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1332.9744
Epoch 142/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1316.0742
Epoch 143/500
2800/2800 [==============================] - 0s 67us/sample - loss: 1314.9390
Epoch 144/500
2800/2800 [==============================] - 0s 69us/sample - loss: 1309.4800
Epoch 145/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1307.4666
Epoch 146/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1304.8010
Epoch 147/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1304.5787
Epoch 148/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1306.1439
Epoch 149/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1297.4934
Epoch 150/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1291.2073
Epoch 151/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1302.2408
Epoch 152/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1286.4586
Epoch 153/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1289.1976
Epoch 154/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1290.9848
Epoch 155/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1285.6917
Epoch 156/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1284.8793
Epoch 157/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1276.9912
Epoch 158/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1282.7271
Epoch 159/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1270.7725
Epoch 160/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1275.8668
Epoch 161/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1269.0021
Epoch 162/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1269.8503
Epoch 163/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1272.6825
Epoch 164/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1256.9528
Epoch 165/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1265.8719
Epoch 166/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1274.9810
Epoch 167/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1251.9277
Epoch 168/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1254.3053
Epoch 169/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1250.6730
Epoch 170/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1247.9147
Epoch 171/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1255.3040
Epoch 172/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1246.4044
Epoch 173/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1248.9053
Epoch 174/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1235.4831
Epoch 175/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1235.7729
Epoch 176/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1231.1521
Epoch 177/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1229.4448
Epoch 178/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1235.0874
Epoch 179/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1236.4246
Epoch 180/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1234.3112
Epoch 181/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1241.1206
Epoch 182/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1228.0536
Epoch 183/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1234.6207
Epoch 184/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1218.8249
Epoch 185/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1227.5636
Epoch 186/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1231.7618
Epoch 187/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1211.7460
Epoch 188/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1214.7899
Epoch 189/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1227.6510
Epoch 190/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1221.4233
Epoch 191/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1208.9139
Epoch 192/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1211.7745
Epoch 193/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1212.7552
Epoch 194/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1225.5838
Epoch 195/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1203.9864
Epoch 196/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1192.8778
Epoch 197/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1196.5854
Epoch 198/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1195.8245
Epoch 199/500
2800/2800 [==============================] - 0s 68us/sample - loss: 1205.7381
Epoch 200/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1232.1300
Epoch 201/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1206.4220
Epoch 202/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1208.0591
Epoch 203/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1184.6516
Epoch 204/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1182.5998
Epoch 205/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1181.9274
Epoch 206/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1192.6584
Epoch 207/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1193.7265
Epoch 208/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1186.7215
Epoch 209/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1199.1962
Epoch 210/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1173.3175
Epoch 211/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1170.1261
Epoch 212/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1168.3255
Epoch 213/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1168.4484
Epoch 214/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1177.4722
Epoch 215/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1167.2033
Epoch 216/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1168.7120
Epoch 217/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1161.4171
Epoch 218/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1179.3264
Epoch 219/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1168.1229
Epoch 220/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1163.9322
Epoch 221/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1168.4568
Epoch 222/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1155.7791
Epoch 223/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1159.8121
Epoch 224/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1144.4758
Epoch 225/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1138.5819
Epoch 226/500
2800/2800 [==============================] - 0s 68us/sample - loss: 1150.9946
Epoch 227/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1142.0983
Epoch 228/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1139.8449
Epoch 229/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1138.2159
Epoch 230/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1132.8268
Epoch 231/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1144.6047
Epoch 232/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1145.6636
Epoch 233/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1131.2519
Epoch 234/500
2800/2800 [==============================] - 0s 67us/sample - loss: 1163.3628
Epoch 235/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1132.9240
Epoch 236/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1123.2252
Epoch 237/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1143.7624
Epoch 238/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1131.8123
Epoch 239/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1120.6352
Epoch 240/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1123.3857
Epoch 241/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1129.1544
Epoch 242/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1121.9190
Epoch 243/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1116.0550
Epoch 244/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1126.5603
Epoch 245/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1115.0059
Epoch 246/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1123.4432
Epoch 247/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1108.1146
Epoch 248/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1109.9104
Epoch 249/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1115.1202
Epoch 250/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1116.5792
Epoch 251/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1109.9456
Epoch 252/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1111.4490
Epoch 253/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1108.3779
Epoch 254/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1104.7345
Epoch 255/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1112.3371
Epoch 256/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1104.8332
Epoch 257/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1114.8357
Epoch 258/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1101.6882
Epoch 259/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1101.5345
Epoch 260/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1103.9248
Epoch 261/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1097.1669
Epoch 262/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1115.2441
Epoch 263/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1112.0905
Epoch 264/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1125.7278
Epoch 265/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1095.7451
Epoch 266/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1097.3662
Epoch 267/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1099.4183
Epoch 268/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1104.0748
Epoch 269/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1091.5516
Epoch 270/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1111.4074
Epoch 271/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1087.1720
Epoch 272/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1100.7326
Epoch 273/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1100.8644
Epoch 274/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1092.4085
Epoch 275/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1096.4041
Epoch 276/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1092.8723
Epoch 277/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1088.8798
Epoch 278/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1085.1037
Epoch 279/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1081.6819
Epoch 280/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1088.1270
Epoch 281/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1085.6152
Epoch 282/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1087.3699
Epoch 283/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1079.3115
Epoch 284/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1075.5549
Epoch 285/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1075.9227
Epoch 286/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1079.0676
Epoch 287/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1083.7895
Epoch 288/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1080.3468
Epoch 289/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1085.7862
Epoch 290/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1078.2151
Epoch 291/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1068.3046
Epoch 292/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1075.6586
Epoch 293/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1071.6740
Epoch 294/500
2800/2800 [==============================] - 0s 66us/sample - loss: 1072.1976
Epoch 295/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1074.4623
Epoch 296/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1077.1401
Epoch 297/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1083.3203
Epoch 298/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1084.7689
Epoch 299/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1080.1851
Epoch 300/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1077.2795
Epoch 301/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1067.7424
Epoch 302/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1068.0804
Epoch 303/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1066.2189
Epoch 304/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1070.1389
Epoch 305/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1068.9136
Epoch 306/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1079.3364
Epoch 307/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1078.0581
Epoch 308/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1100.6148
Epoch 309/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1065.8257
Epoch 310/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1066.7104
Epoch 311/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1060.9756
Epoch 312/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1064.4785
Epoch 313/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1062.8892
Epoch 314/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1076.1493
Epoch 315/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1075.2968
Epoch 316/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1062.3199
Epoch 317/500
2800/2800 [==============================] - 0s 57us/sample - loss: 1064.0583
Epoch 318/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1073.0384
Epoch 319/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1060.9311
Epoch 320/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1060.9842
Epoch 321/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1065.6132
Epoch 322/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1058.8536
Epoch 323/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1056.9967
Epoch 324/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1051.3154
Epoch 325/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1052.8784
Epoch 326/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1042.9993
Epoch 327/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1044.8463
Epoch 328/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1051.4864
Epoch 329/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1043.7003
Epoch 330/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1051.8080
Epoch 331/500
2800/2800 [==============================] - 0s 53us/sample - loss: 1054.5417
Epoch 332/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1057.1723
Epoch 333/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1053.7270
Epoch 334/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1060.2016
Epoch 335/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1049.5115
Epoch 336/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1041.7271
Epoch 337/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1031.7273
Epoch 338/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1044.8150
Epoch 339/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1046.0760
Epoch 340/500
2800/2800 [==============================] - 0s 55us/sample - loss: 1049.0824
Epoch 341/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1040.2875
Epoch 342/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1030.4299
Epoch 343/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1043.7071
Epoch 344/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1034.6867
Epoch 345/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1037.4047
Epoch 346/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1039.7557
Epoch 347/500
2800/2800 [==============================] - 0s 56us/sample - loss: 1024.1343
Epoch 348/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1027.6918
Epoch 349/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1018.4648
Epoch 350/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1013.8253
Epoch 351/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1013.7689
Epoch 352/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1023.0029
Epoch 353/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1017.0246
Epoch 354/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1024.0550
Epoch 355/500
2800/2800 [==============================] - 0s 67us/sample - loss: 1009.9415
Epoch 356/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1003.8997
Epoch 357/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1006.5315
Epoch 358/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1016.5423
Epoch 359/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1010.5035
Epoch 360/500
2800/2800 [==============================] - 0s 64us/sample - loss: 1002.4500
Epoch 361/500
2800/2800 [==============================] - 0s 58us/sample - loss: 1015.2172
Epoch 362/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1007.1971
Epoch 363/500
2800/2800 [==============================] - 0s 61us/sample - loss: 1006.8336
Epoch 364/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1007.2265
Epoch 365/500
2800/2800 [==============================] - 0s 59us/sample - loss: 1007.9150
Epoch 366/500
2800/2800 [==============================] - 0s 62us/sample - loss: 1006.6157
Epoch 367/500
2800/2800 [==============================] - 0s 63us/sample - loss: 997.9993
Epoch 368/500
2800/2800 [==============================] - 0s 63us/sample - loss: 1010.8255
Epoch 369/500
2800/2800 [==============================] - 0s 60us/sample - loss: 1010.5538
Epoch 370/500
2800/2800 [==============================] - 0s 61us/sample - loss: 997.4726
Epoch 371/500
2800/2800 [==============================] - 0s 61us/sample - loss: 997.9302
Epoch 372/500
2800/2800 [==============================] - 0s 58us/sample - loss: 988.6962
Epoch 373/500
2800/2800 [==============================] - 0s 63us/sample - loss: 995.6228
Epoch 374/500
2800/2800 [==============================] - 0s 62us/sample - loss: 989.1748
Epoch 375/500
2800/2800 [==============================] - 0s 60us/sample - loss: 992.7116
Epoch 376/500
2800/2800 [==============================] - 0s 64us/sample - loss: 987.3237
Epoch 377/500
2800/2800 [==============================] - 0s 63us/sample - loss: 992.6085
Epoch 378/500
2800/2800 [==============================] - 0s 62us/sample - loss: 994.6691
Epoch 379/500
2800/2800 [==============================] - 0s 65us/sample - loss: 1008.5694
Epoch 380/500
2800/2800 [==============================] - 0s 63us/sample - loss: 986.9961
Epoch 381/500
2800/2800 [==============================] - 0s 61us/sample - loss: 985.3143
Epoch 382/500
2800/2800 [==============================] - 0s 64us/sample - loss: 980.2779
Epoch 383/500
2800/2800 [==============================] - 0s 64us/sample - loss: 983.7935
Epoch 384/500
2800/2800 [==============================] - 0s 64us/sample - loss: 982.4315
Epoch 385/500
2800/2800 [==============================] - 0s 60us/sample - loss: 978.9523
Epoch 386/500
2800/2800 [==============================] - 0s 61us/sample - loss: 972.5764
Epoch 387/500
2800/2800 [==============================] - 0s 62us/sample - loss: 974.3637
Epoch 388/500
2800/2800 [==============================] - 0s 59us/sample - loss: 984.3422
Epoch 389/500
2800/2800 [==============================] - 0s 61us/sample - loss: 980.3249
Epoch 390/500
2800/2800 [==============================] - 0s 66us/sample - loss: 976.2236
Epoch 391/500
2800/2800 [==============================] - 0s 61us/sample - loss: 981.1634
Epoch 392/500
2800/2800 [==============================] - 0s 63us/sample - loss: 969.4153
Epoch 393/500
2800/2800 [==============================] - 0s 60us/sample - loss: 981.6107
Epoch 394/500
2800/2800 [==============================] - 0s 61us/sample - loss: 968.6300
Epoch 395/500
2800/2800 [==============================] - 0s 64us/sample - loss: 979.7458
Epoch 396/500
2800/2800 [==============================] - 0s 68us/sample - loss: 962.2366
Epoch 397/500
2800/2800 [==============================] - 0s 62us/sample - loss: 981.2674
Epoch 398/500
2800/2800 [==============================] - 0s 64us/sample - loss: 978.5301
Epoch 399/500
2800/2800 [==============================] - 0s 67us/sample - loss: 969.2405
Epoch 400/500
2800/2800 [==============================] - 0s 62us/sample - loss: 961.5628
Epoch 401/500
2800/2800 [==============================] - 0s 66us/sample - loss: 970.1200
Epoch 402/500
2800/2800 [==============================] - 0s 65us/sample - loss: 977.9712
Epoch 403/500
2800/2800 [==============================] - 0s 64us/sample - loss: 962.8104
Epoch 404/500
2800/2800 [==============================] - 0s 61us/sample - loss: 969.9754
Epoch 405/500
2800/2800 [==============================] - 0s 67us/sample - loss: 968.4667
Epoch 406/500
2800/2800 [==============================] - 0s 69us/sample - loss: 956.4312
Epoch 407/500
2800/2800 [==============================] - 0s 65us/sample - loss: 967.2131
Epoch 408/500
2800/2800 [==============================] - 0s 63us/sample - loss: 972.8434
Epoch 409/500
2800/2800 [==============================] - 0s 60us/sample - loss: 970.3013
Epoch 410/500
2800/2800 [==============================] - 0s 64us/sample - loss: 956.3091
Epoch 411/500
2800/2800 [==============================] - 0s 65us/sample - loss: 955.3719
Epoch 412/500
2800/2800 [==============================] - 0s 75us/sample - loss: 968.8319
Epoch 413/500
2800/2800 [==============================] - 0s 63us/sample - loss: 964.6352
Epoch 414/500
2800/2800 [==============================] - 0s 62us/sample - loss: 956.8947
Epoch 415/500
2800/2800 [==============================] - 0s 65us/sample - loss: 969.1176
Epoch 416/500
2800/2800 [==============================] - 0s 66us/sample - loss: 955.3204
Epoch 417/500
2800/2800 [==============================] - 0s 66us/sample - loss: 947.9882
Epoch 418/500
2800/2800 [==============================] - 0s 64us/sample - loss: 963.2110
Epoch 419/500
2800/2800 [==============================] - 0s 64us/sample - loss: 956.1462
Epoch 420/500
2800/2800 [==============================] - 0s 64us/sample - loss: 967.4477
Epoch 421/500
2800/2800 [==============================] - 0s 65us/sample - loss: 958.8824
Epoch 422/500
2800/2800 [==============================] - 0s 64us/sample - loss: 955.9722
Epoch 423/500
2800/2800 [==============================] - 0s 62us/sample - loss: 943.9016
Epoch 424/500
2800/2800 [==============================] - 0s 67us/sample - loss: 946.2346
Epoch 425/500
2800/2800 [==============================] - 0s 67us/sample - loss: 957.0402
Epoch 426/500
2800/2800 [==============================] - 0s 60us/sample - loss: 947.0486
Epoch 427/500
2800/2800 [==============================] - 0s 61us/sample - loss: 946.5983
Epoch 428/500
2800/2800 [==============================] - 0s 65us/sample - loss: 953.2834
Epoch 429/500
2800/2800 [==============================] - 0s 65us/sample - loss: 945.5103
Epoch 430/500
2800/2800 [==============================] - 0s 62us/sample - loss: 964.1069
Epoch 431/500
2800/2800 [==============================] - 0s 59us/sample - loss: 949.8702
Epoch 432/500
2800/2800 [==============================] - 0s 59us/sample - loss: 946.5142
Epoch 433/500
2800/2800 [==============================] - 0s 62us/sample - loss: 965.2561
Epoch 434/500
2800/2800 [==============================] - 0s 62us/sample - loss: 952.1427
Epoch 435/500
2800/2800 [==============================] - 0s 61us/sample - loss: 954.9595
Epoch 436/500
2800/2800 [==============================] - 0s 66us/sample - loss: 937.4476
Epoch 437/500
2800/2800 [==============================] - 0s 71us/sample - loss: 945.1045
Epoch 438/500
2800/2800 [==============================] - 0s 63us/sample - loss: 939.1078
Epoch 439/500
2800/2800 [==============================] - 0s 61us/sample - loss: 939.0538
Epoch 440/500
2800/2800 [==============================] - 0s 65us/sample - loss: 947.7939
Epoch 441/500
2800/2800 [==============================] - 0s 64us/sample - loss: 942.8379
Epoch 442/500
2800/2800 [==============================] - 0s 61us/sample - loss: 943.3696
Epoch 443/500
2800/2800 [==============================] - 0s 61us/sample - loss: 942.1774
Epoch 444/500
2800/2800 [==============================] - 0s 61us/sample - loss: 939.1387
Epoch 445/500
2800/2800 [==============================] - 0s 59us/sample - loss: 948.7494
Epoch 446/500
2800/2800 [==============================] - 0s 65us/sample - loss: 943.7548
Epoch 447/500
2800/2800 [==============================] - 0s 62us/sample - loss: 932.7881
Epoch 448/500
2800/2800 [==============================] - 0s 61us/sample - loss: 939.7122
Epoch 449/500
2800/2800 [==============================] - 0s 56us/sample - loss: 932.3667
Epoch 450/500
2800/2800 [==============================] - 0s 64us/sample - loss: 960.0179
Epoch 451/500
2800/2800 [==============================] - 0s 62us/sample - loss: 940.8460
Epoch 452/500
2800/2800 [==============================] - 0s 61us/sample - loss: 949.1144
Epoch 453/500
2800/2800 [==============================] - 0s 65us/sample - loss: 936.2326
Epoch 454/500
2800/2800 [==============================] - 0s 60us/sample - loss: 930.4036
Epoch 455/500
2800/2800 [==============================] - 0s 62us/sample - loss: 938.6048
Epoch 456/500
2800/2800 [==============================] - 0s 62us/sample - loss: 937.3702
Epoch 457/500
2800/2800 [==============================] - 0s 63us/sample - loss: 938.8040
Epoch 458/500
2800/2800 [==============================] - 0s 61us/sample - loss: 959.1447
Epoch 459/500
2800/2800 [==============================] - 0s 59us/sample - loss: 929.1206
Epoch 460/500
2800/2800 [==============================] - 0s 60us/sample - loss: 934.5046
Epoch 461/500
2800/2800 [==============================] - 0s 61us/sample - loss: 937.0909
Epoch 462/500
2800/2800 [==============================] - 0s 65us/sample - loss: 932.0309
Epoch 463/500
2800/2800 [==============================] - 0s 56us/sample - loss: 946.6648
Epoch 464/500
2800/2800 [==============================] - 0s 63us/sample - loss: 926.3534
Epoch 465/500
2800/2800 [==============================] - 0s 60us/sample - loss: 926.0744
Epoch 466/500
2800/2800 [==============================] - 0s 59us/sample - loss: 928.1909
Epoch 467/500
2800/2800 [==============================] - 0s 64us/sample - loss: 921.8013
Epoch 468/500
2800/2800 [==============================] - 0s 62us/sample - loss: 942.3816
Epoch 469/500
2800/2800 [==============================] - 0s 75us/sample - loss: 928.0069
Epoch 470/500
2800/2800 [==============================] - 0s 64us/sample - loss: 923.5392
Epoch 471/500
2800/2800 [==============================] - 0s 60us/sample - loss: 929.2638
Epoch 472/500
2800/2800 [==============================] - 0s 63us/sample - loss: 933.0489
Epoch 473/500
2800/2800 [==============================] - 0s 62us/sample - loss: 926.7229
Epoch 474/500
2800/2800 [==============================] - 0s 61us/sample - loss: 923.6627
Epoch 475/500
2800/2800 [==============================] - 0s 58us/sample - loss: 924.1287
Epoch 476/500
2800/2800 [==============================] - 0s 58us/sample - loss: 935.6834
Epoch 477/500
2800/2800 [==============================] - 0s 59us/sample - loss: 937.3834
Epoch 478/500
2800/2800 [==============================] - 0s 62us/sample - loss: 934.3209
Epoch 479/500
2800/2800 [==============================] - 0s 58us/sample - loss: 919.9716
Epoch 480/500
2800/2800 [==============================] - 0s 67us/sample - loss: 943.3614
Epoch 481/500
2800/2800 [==============================] - 0s 59us/sample - loss: 923.1147
Epoch 482/500
2800/2800 [==============================] - 0s 61us/sample - loss: 921.4427
Epoch 483/500
2800/2800 [==============================] - 0s 59us/sample - loss: 913.0294
Epoch 484/500
2800/2800 [==============================] - 0s 63us/sample - loss: 915.4022
Epoch 485/500
2800/2800 [==============================] - 0s 56us/sample - loss: 922.7092
Epoch 486/500
2800/2800 [==============================] - 0s 60us/sample - loss: 932.7826
Epoch 487/500
2800/2800 [==============================] - 0s 60us/sample - loss: 919.0049
Epoch 488/500
2800/2800 [==============================] - 0s 61us/sample - loss: 913.9195
Epoch 489/500
2800/2800 [==============================] - 0s 62us/sample - loss: 928.3425
Epoch 490/500
2800/2800 [==============================] - 0s 61us/sample - loss: 939.2139
Epoch 491/500
2800/2800 [==============================] - 0s 59us/sample - loss: 929.5971
Epoch 492/500
2800/2800 [==============================] - 0s 64us/sample - loss: 912.6969
Epoch 493/500
2800/2800 [==============================] - 0s 61us/sample - loss: 912.2640
Epoch 494/500
2800/2800 [==============================] - 0s 63us/sample - loss: 930.4466
Epoch 495/500
2800/2800 [==============================] - 0s 62us/sample - loss: 920.4040
Epoch 496/500
2800/2800 [==============================] - 0s 62us/sample - loss: 914.7307
Epoch 497/500
2800/2800 [==============================] - 0s 63us/sample - loss: 913.2255
Epoch 498/500
2800/2800 [==============================] - 0s 61us/sample - loss: 924.5638
Epoch 499/500
2800/2800 [==============================] - 0s 63us/sample - loss: 906.7920
Epoch 500/500
2800/2800 [==============================] - 0s 58us/sample - loss: 916.3150





<tensorflow.python.keras.callbacks.History at 0x7f6dd240abe0>

preds = model.predict(x_test)

# Results

rms = np.sqrt(np.mean(np.power((np.array(y_test)-np.array(preds)),2)))

print(rms)

600.8828804286148

test['Predictions'] = 0
test['Predictions'] = preds


plt.plot(train['Adj. Close'])
plt.plot(test[['Adj. Close', 'Predictions']])

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy

/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy






[<matplotlib.lines.Line2D at 0x7f6dd20c25f8>,
 <matplotlib.lines.Line2D at 0x7f6dd20c2828>]

egsavage/Stock-Price-Prediction

Overview

Talk about market efficiency and the paradox of passive investors and active investors

Table of Contents

The Stocker Module

Evaluation of Predictions

Adjusting the Changepoint Priors

Evaluating Refined Model

Testing Our Luck in the Stock Market

Preparing Data for More Machine Learning

Moving Average

Simple Linear Regression

k-Nearest Neighbours

Multilayer Perceptron