TDDE01 is a Machine Learning course taught at Linköping University, Sweden. Attended Fall 2022. The course includes three labs, with three assigments each, which goes through most of the basic statistical models used in machine learning, as well as introducing kernels, SVMs, and neural networks.
Lab 1 - k-NN, Regression, Basis Function Expansion
Assignment 1 implements a K-nearest neighbors algorithm to predict handwritten digits ranging from 0-9.
Assignment 2 uses Linear Regression and Ridge Regression to predict Parkinson's disease symptom score.
Assignment 3 implements Logistic regression to predict diabetes. The results are compared to using basis function expansion.
Lab 2 - LASSO Regression, Decision trees, PCA
Assignment 1 uses LASSO Regression to investigate whether a near infrared absorbance spectrum can be used to predict fat content in samples of meat.
Assignment 2 implements Decision trees to determine the success of a bank marketing campaign.
Assignment 3 implements a Principcal Component Analysis to research correlations of violent crimes and a number of features such as economic situation etc.
Lab 3 - Kernel Methods, Support Vector Machines, Neural Networks
Assignment 1 uses kernel methods to predict hourly temperatures for a date and place in Sweden. Data provided by the Swedish Meteorological and Hydrological Institute (SMHI).
Assignment 2 implements an SVM model to classify spam messages.
Assignment 3 trains a Neural Network to learn the trigonometric sine function.