Mathematical Sciences Department Colloquium: Sajal Chakroborty, WPI
11:00 am to 11:50 am
Tuesday, October 4th 2024
11:00 – 11:50 AM
Stratton Hall 202
Title:Nonparametric Statistical Tools from Potential Theory with a Focus on Outlier Detection
Abstract: In Data Science, Statistical tools are essential for analyzing and interpreting data, identifying hidden patterns, and predicting future trends. However, the construction of Statistical tools for high-dimensional spaces (HDS) is not straightforward. One of the obstacles is that we cannot rank data naturally in a space with dimensions d≥2d≥2. Moreover, statistical computations in HDS, such as computing cumulative distribution function is much more complex than their one-dimensional counterpart. In this research, we develop some nonparametric statistical tools in Rd, d≥2ℝd, d≥2 from Potential theory with a focus on outlier detection.