Bayesian Online Change Point Detection in Finance

Abstract It is quite common that the structure of a time series changes abruptly. Identifying these change points and describing model in segments between an important task financial analysis. Change point detection identification abrupt generative parameters sequential data. In application areas such as finance, online rather than offline mostly required, due to their use predictive tasks, possibly embedded automatic trading systems. However, complex data generation processes makes this challenging endeavor. This paper concerned with using Bayesian setting. To end, posterior probability at specific proposed some procedures are presented for selecting priors estimation parameters. Applications simulated given. Finally, conclusions proposed.