Estimation of Exogenous and Endogenous Factors in Stock Market Transactions and Quantification of Market States
Point process analysis considers the sequence of time stamps at which each event occurs. We analyzed the point processes of transactions on the Tokyo Stock Exchange (TSE) to quantify the state of the market. In this analysis, we improved an algorithm to infer the time evolution of the strength of exogenous and endogenous factors driving transactions and applied it to transaction point processes for various stocks on the TSE. We quantified the impact of external events (such as policy change announcements) on transactions and found that stock responses to exogenous and endogenous factors vary according to characteristics such as market capitalization.
Construction of Early Warning Signals of Market Instability Based on Dynamical Network Marker (DNM) Theory
DNM theory provides a method for constructing early warning signals based on dynamical systems theory. Dynamical Network Biomarker (DNB) theory, developed to detect pre-disease states within the same framework as DNM theory, has been extensively studied, targeting the Ultra-Early Precision Medicine. In our project, we apply DNM theory to a system of interacting elements in a financial market, with each market participant treated as an element. We aim to detect early warning signals of market instability by analyzing a large dataset of order placements and transactions using DNM theory.
