NahidEbrahimian/Machine-Learning

Machine Learning algorithms Implementation from Scratch

Jupyter Notebook

Machine-Learning

01- Introductory-Machine-Learning

In this tutorial:

Mathematics and study statistics, and how to calculate important numbers based on data set

Reference site: w3schools

03- KNN Classifier

Implementation of KNeighbors Classifier
Evaluating model using Normalized random data

04- OCR of Hand-written using KNN

Building a basic OCR (Optical Character Recognition) application algorithm using KNeighbors Classifier
Evaluating model using Hand-written Digits related to Mnist numbers

Reference sites: opencv

Mnist, Hand-written numbers image: Mnist numbers

05- KNN on Iris Dataset

Apply Knn classifier with different values of k on Iris Dataset and plot test accuracy
Calculate and plot Confusion Matrix for predicted values with k = 5

06- Adaline Regression

Adaline on random generated data for weight and height of humans
linregress of scipy package - plot slope

07- Adaline Classification

Adaline on random generated data for hair length of men and women

08- AdalineRegressor on Boston Dataset

09- AdalineClassifier on Iris

Select two classes of Iris Dataset and Apply Adaline Classifier
Calculate accuracy
Comparison accuracy of Adaline Classifier with KNN on Select classes

10- Perceptron-Classification

Implementation of perceptron class, evaluate and predict function
Perceptron-Classification on linear data

11- Perceptron-Regression

Implementation of perceptron class, evaluate and predict function
Preprocessing data using pandas library
Perceptron-Regression on weatherHistory dataset

12-1 MLP on Titanic Dataset

MLP on Titanic Dataset, sequential model - tensorflow and keras
12-2, 12-3, 12-4: Comparison accuracy of MLP with KNN, Adaline and perceptron on Titanic Dataset
Result:

MLP accuracy = 96%

Knn accuracy = 65%

Adaline accuracy = 86%

perceptron accuracy = 34%

Dataset link: titanic

13-MLP on House Prices Dataset

Result on wandb: MLP-HousePrices
Dataset link: house-prices

14- WeatherPrediction_Regression

Train Neural Network on weather-dataset using tensorflow and keras

Dataset:

Dataset link: weather-dataset

Loss on test data: 3.0455

15- learning-rate-schedules

Learning Rate schedules in Practice