Project for Machine Learning and Pattern Recognition course regarding detection of quality of wine.
Problem was simplified to binary classification. The goal of the project was to classify the Test.txt set of wines based on Training.txt given 10 different parameters of each sample.
The project aimed at learning and utilizing knowledge od basics of Machine Learning without using in-built and external Machine Learning libraries (limitation to numpy, scipy and matplotlib). All the methods used in the project where written using libraries mentioned earlier. In the project I present methods like:
- PCA
- LDA
- various variations of Gaussian Models
- Logistic Regression
- Support Vector Machines
- Gaussian Mixture Models
- model evaluation methods
Further details regarding theoretical aspects and outcome of the project are attached in report.pdf.