BenyaminZojaji/Machine_Learning

Data Analysis and ML in Python.

machine_learning

Assignment 31

• Implement the Machine Learning and Matplotlib training section of the w3schools website.
  • Matplotlib
  • Machine Learning

---

Assignment 32

3d Scatter

• A set of random numbers with three properties of length, width and weight for balloons and melons.
• Display the generated data in three dimensions.

Fruit Visualization

• python visualization (case) fruit classification (banana and Apple war).

---

Assignment 33

Know Your Metrics

• Analysis an Online Retail data set to find out the problem in April.

---

Assignment 34

Covid

• The eight countries with the highest number of covid cases in the fourth month of 2020.

• Draw a graph that shows the mortality rate in relation to the number of cases in Iran on different days.

KNN

• Write KNN(K Nearest Neighbors) algorithm from scratch then compare with sklearn KNeighborsClassifier.
• Working on weight-height Dataset on kaggle.

KNN OCR

• use kNN to build a basic OCR (Optical Character Recognition) application.
• In this case we work on mnist(Modified National Institute of Standards and Technology dataset) numbers.
• result -> accuracy: 91.76

---

Assignment 35

Nemo

• Train kNN algorithm with Clownfish¹ image and test it on another Clownfish image.
• written in Python using opencv, matplotlib.

Iris EDA²

• Doing kNN algorithm on sckit-learn Iris dataset with different k and plot the accuracy.

Abalone EDA²

• Doing kNN algorithm on Abalone Dataset and obtain the accuracy of the algorithm.
• NOTE: In this exercise, I solved the problem in the form of classification.
• NOTE: This problem is solved in form of regression in Assignment 37

---

Assignment 36

Iris

• Drawing confusion matrix for the iris problem from last assignment.

LLS (Linear Least-Squares)

• Creating continuous random data for students' study hours and their grades.

• Obtaining line slope by LLS methods. (formula and scipy library)

• Draw both of them in one figure.

LLS - 2 independent

• Implementing the LLS method on the Boston dataset from the Scikit-learn library.
• This database offers thirteen features per house, I chose 'CRIM' and 'TAX'.
• Lets scatter the data.

• Apply the LLS method on data and get the predicted-surface and plot it.

• Lets plot data and predicted-surface in one figure.

• Lets generate the three different figures from different views to see better.

---

Assignment 37

DigiKala

• EDA² on DigiKala's order dataset.
• Number of orders per month:

• Separate customers by city:

Abalone Regression

• Doing LLS algorithm on Abalone Dataset and calculate MAE, MSE, Huber, and Hinge Errors.

QR Code reader

• QR Code reader written in python using opencv detectAndDecode method that can decode QR Codes and Barcodes.

---

Assignment 38

Perceptrone

---

Assignment 39

Weather Forecast: (Linear Regression)

• The climate of a particular city is recorded every hour during different years. So, weather information is recorded 24 times per day.
• Draw an output table on a chart.
• Teach a linear model³ on the above data using the perceptron algorithm

• Plot Loss and R2-score diagrams.

---

Assignment 40

Titanic⁴

Algorithm	KNN	Perceptron	MLP (Multi Layer Perceptron)
Accuracy	72.66%	43.16%	92.57%

---

Assignment 41

MLP-Weather

• solve Assignment39 with Multi Layer Perceptron.

epoch	5	10	100	800
Loss	1.10	0.69	0.17	0.13

Footnotes

1. Clownfish in wikipedia ↩

2. EDA stands for Exploratory Data Analysis ↩ ↩² ↩³

3. I am aware of linear model can not fit the data perfectly. This assignment solved with mlp in Assignment 41. ↩

4. Titanic Dataset ↩