Module Project For Machine Learning, looking at the relationship between sprectrum reading and viral load severity. Using both suppervised and unsupervised approaches for viral load prediction. Then, classifying viruses using different algorithms, achieving ~90% accuracy.
-
Feature Engineering: PCA, PLS, Lasso
-
Regression: Linear, polynomial, Neural Network
-
Classification: Gaussian Naive Bayes (NB), Logistic Regression, K Nearest Neighbours (KNN), Linear Discriminate Analysis (LDA) and Quadratic Discriminate Analysis (QDA), Partial Least Squares Discriminate Analysis
-
Clustering: Kmeans, AGC clustering, DBSCAN
-
Report PDF attached.