Seminar | March 6 | 3-4 p.m. | 540AB Cory Hall
Ayush Pandey, Ph.D. Candidate, CalTech
Electrical Engineering and Computer Sciences (EECS)
In this teaching demonstration, I will present the Singular Value Decomposition (SVD) and its diverse applications in data science problems. I will build the principles underlying dimensionality reduction in large datasets by drawing on students prior background in linear algebra (eigenvalues, matrix operations). I will formally describe the singular value decomposition and how it forms the basis of some of the most common modeling techniques in data science such as least squares and PCA. The analytical proofs will be followed up with a brief introduction to two application examples from (1) image processing and (2) biological gene expression analysis.
CA, kcisco@berkeley.edu, 510-642-0954
Katie Cisco, kcisco@berkeley.edu, 510-642-0954