Machine learning algorithms allow us to induce rules based on empirical training data. The course establishes statistical and algebraic foundations essential to kernel-based learning approach, for vectorial and nonvectorial data analysis. More specifically, it covers: (i) aptive techniques for feature selection and dimension reduction PCA, SODA and DCA; (ii) unsupervised cluster discovery: K-means, SOM, kernel K-means, kernel SOM, and hierarchical clustering; and (iii) supervised learning algorithms: Fisher discriminate analysis (FDA), kernel ridge regressor (KRR), support vector machines (SVM), and ridge-SVM.
Sample reading list:
S.Y. Kung, Kernal Methods and Machine Learning (Cambridge Press, 2014)
Not Open to Freshmen.
The grade will be based on course projects.
|42693||S01||9:30 am - 10:50 am||M W||Engineering Quad A-Wing A224||Enrolled:10 Limit:16|