Machine learning of structural relationships among variables from empirical data. Decision theory, Bayesian methods. Classification: naïve Bayes, linear discriminant analysis, support vector machines (SVM), boosting. Regression: leastsquares, regularization methods, logistic regression. Clustering using kmeans and EM algorithms. Model selection: bias-variance tradeoff, crossvalidation, over-fitting. Feature selection and dimensionality reduction methods including PCA, ICA, MDS. Kernel methods. Other topics may be covered as time permits.
Prerequisites: Ma 223, Ma 224; either ECE 211, ChE 352 or ME 251
Open to all students.
Course Code: ECE 414