This repository contains some interesting python and machine learning projects that I have worked on during my studies.

Python and Deep Learning Projects

AI

Neural Style Transfer

This project involves using convolutional neural networks to perform image style transfer.

• main.py: contains the main part of the code where a pre-trained VGG-19 model (using TensorFlow).
• creating-model: involves creating a neural network and training it, with the goal of using this model instead of VGG-19 and benchmarking it.

Machine Learning in R

This folder contains multiple R Markdown reports where I have worked on the following machine learning techniques:

• Bayesian unsupervised clustering
• Linear regression applied to a problem of data mining
• Principal components analysis and principle components regression applied to predicting geographic location based on genetic traits
• Benchmarking multiple classification techniques

Rule-Based Mining

This involves processing raw data, generating frequent itemsets, and filtering association rules based on certain criteria such as minimum support, confidence, and lift. These rules can be used to gain insights into the underlying patterns and relationships between items in the dataset.

Diverse Python Projects

Kinect Project

This project involves capturing a 3D surface (face) using a Kinect, and using B-Spline curves to modify and reconstruct a part of the 3D surface.

Predator-Prey Simulation

This involves implementing the Lotka-Volterra model to simulate predator-prey interactions.

Bezier Intersect

Detecting intersection between two Bezier curves.

De Boor Cox

Using De Boor Cox algorithms to plot B-Spline curves.

Minimal Distance Between Two Points in 2D Plan

Finding efficiently a pair of points in a 2D plan. We explore 3 methods and benchmark them.

Introductory Research Project: Optimization of Immunization

This involves modeling the population as a scale-free network, simulating the propagation of an epidemic in the network, and using different vaccination techniques to determine the threshold for percolation.

Approximate PI

Using Monte Carlo approximation to approximate the value of pi and visualize the approximation process as a gif.