Nonlinear Statistical Modeling for High-Dimensional and Small Sample Data

Explore nonlinear statistical modeling techniques for high-dimensional and small sample data in this 57-minute seminar presented by Dr. Makoto Yamada, a renowned expert in machine learning and statistical science. Delve into advanced feature selection methods designed to handle large-scale, ultra-high-dimensional datasets with over a million features. Learn about innovative approaches such as HSIC Lasso, kernel PSIs, and feature selection networks, which offer powerful solutions for nonlinear feature selection in challenging data environments. Gain insights into the applications of these techniques across various fields, including biology, natural language processing, and computer vision. Discover how these cutting-edge methods can be applied to real-world problems like gene selection from microarray data, document categorization, and prosthesis control.