Topological Approach for Detection of Structural Breakpoints in Philippine Stock Price Data Surrounding the COVID-19 Pandemic

Topological data analysis (TDA) is a novel but rigorous framework used to identify the underlying shape and structure of a huge data set. In this study, we demonstrated how TDA can be useful in analyzing and predicting stock market movement. The study collected time-series data from the Philippine Stock Exchange of companies that were heavily affected by lockdown measures implemented during the COVID-19 pandemic and those that thrived. The study found that during market crashes, persistence diagrams revealed a presence of persistence weakening phenomenon. Moreover, by analyzing the stock price movement given an interval of time, the persistence landscape graphs produced high points or bumps during a market drop, which corresponds to structural breakpoints in the time series. Indeed, persistent homology can be used to predict early warning signs of probable market crashes.