Data Science Seminar
Date: June 18, 2021 (Friday)
Time: 2:00 pm - 3:00 pm
Universal Consistency of the k-NN Rule: A Review
Chennai Mathematical Institute.
The k-NN rule is the most important learning rule in statistical machine learning whose consistency properties (that is, error convergence) help in developing the framework of other learning algorithms. In this talk, I will discuss the universal consistency of the k-NN rule in various metric spaces. In addition, we will analyze the combination of k-NN rule and random projection (a dimension reduction method) from a theoretical angle.
Short Bio: Sushma has recently joined CMI as a post-doctoral fellow. She has obtained her M.Sc. in mathematics from IIT Hyderabad and further her Ph.D. from the Kyoto University, Japan. Her thesis was based on random matrices and statistical machine learning. She is interested in studying the consistency properties of traditional machine learning rules when combined with dimensionality reduction methods.