ST310 Machine Learning

Course work completed as part of the ST310 module taught at the LSE.

Topics covered:

  • Regression : Linear, Polynomial;

  • General Linear Models (GLM);

  • Models with high levels of flexibility : eg. Locally Estimated Scatterplot Smoothing (LOESS);

  • Models with low levels of flexibility : Least Absolute Shrinkage and Selection Operator (LASSO) and Ridge Regression;

  • Prediction on non-numeric qualitative outcomes : Classification, Logistic Regression;

  • Optimisation techniques involving Gradient descent, Coordinate descent, Stochastic/Random descent;

  • Generalisation and overfitting problems: In Distribution (Overfitting to variance) & Out Of Distribution (Overfitting to bias).

Book reference: ISLR https://www.statlearning.com/