This repository contains my work for this course at my university.
Original course title: Metody inteligencji obliczeniowej w analizie danych
- Neural Network - folder
MLP- written from scratch in numpy
- has many features such as:
moment,rms prop,softmaxand classification with F1 score- different activation functions,
- regularization L1 and L2,
- early stopping and others,
- Self Organizing Map - folder
SOM- rectangular and hexagonal maps
- neighborhood functions: mexican hat and gaussian
- decreasing learning rate
- Genetic Algorithm - folder
AE- minimize a simple function
- cutting problem - packing rectangles into a circle
- training a neural network given it's architecture