Machine Learning Tasks - repository

This repository contains solution of some tasks from book An introduction to Statistical Learning and tasks suggested during machine learning classes during medical computer science course at AGH University of Science and Technology.

Scope of classes

Regression

• regression model parameters estimation
• MSE estimation and other statistics
• p-value testing for model parameters
• Regression models with torch and SGD optimization

Regression Tasks

Classification

• classifiers: Logistic Regression, LDA, QDA, Gaussian Bayes and KNN
• MNIST digits data set classification (with HOG)
• well known statistics calculation (such as ROC area)
• model parameters optimization with Optuna

Classification Tasks

ResamplingMethods

• Monte Carlo Methods
• Bootstrap Methods
• Permutation Methods (for correlated and uncorrelated data)
• Cross-Validation

Resampling Methods

Tree Methods

• Simple Trees
• Tree Pruning
• Nodes Purity Analysis
• Feature Importance
• Drawing Tree
• Random Forest

Classification Trees with analysis and pruning
Regression Trees with Random Forest Regressor Params Analysis

Project - Exam Score Prediction

This Project Includes following steps:

• Feature Engineering and Data Exploration
• Model Selection
• Statistical Testing of Piplines
• Coefficients Analysis of Linear Model
• Analysis of Random Forest and Simple Decision Tree
• Final Model with example Observation Prediction

Project Exam Score Prediction

Special Thanks

Special Thanks for Professor Zbisław Tabor for running such a good classes.