This repository contains popular Machine Learning algorithms made from scratch, only using NumPy.
- For given objective functions Newton's method makes optimazation and finds maximum of function.
- Implementation of evolutionary algorithm without breeding with tournament selection and elite succession to find maximum and minimum of given functions.
- Implementation of minimax algorithm (without pruning) and use of it in Tic-Tac-Toe game.
- Implementation of model, method teaching linear SVM, optimazation of hiperparameters using gradient descent and classification on IRIS dataset.
- Implementation of Naive Bayes Classifier, use of it for Wine dataset classification and checking its correctness with k-fold cross validation.
- Implementation of neural network as multilayer perceptron model with backpropagation as method teaching him and use of it for Mnist dataset classification.
- Implementation of reinforcement learning through Q-Learning algorithm with OpenAI gym compatibility and its test on FrozenLake8x8-v0 environment.
- Python 3.9.5.
- Python libraries and packages that you can find in requirements.txt file.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent