Speaker:  Dr. Arnab K Laha

Indian Institute of Management, Ahmedabad, India

Time Series Clustering and Classification using Topological Data Analysis with applications to Finance

In this talk, we discuss new methods for time series classification and clustering based on techniques of Topological Data Analysis (TDA). We present a new clustering method SOM-TDA using which we examine the topological similarities and dissimilarities of some well-known time-series models. A new classification method RF-TDA based on TDA is introduced and is used to examine if the topological features can be used to distinguish between time series models. The RF-TDA method is then used for classification of stock price time series of different sector stocks traded at the National Stock Exchange (NSE), India. The results imply that the topological features of the time series of stock prices in the different sectors are not identical and have distinctive features that can be discerned through the use of TDA.

This work is joint with Sourav Majumdar.