Model Introduced SPRT for Structural Change Detection of Time Series (I) -- Formulation --

Authors
Yoshihide Koyama, Tetsuo Hattori, Hiromichi Kawano
Corresponding Author
Yoshihide Koyama
Available Online 30 June 2014.
DOI
https://doi.org/10.2991/jrnal.2014.1.1.11How to use a DOI?
Keywords
Time series, Change detection, SPRT (Sequential Probability Ratio Test), Hidden Markov Model
Abstract
Previously, we have proposed a method applying Sequential Probability Ratio Test (SPRT) to the structural change detection problem of ongoing time series data. In this paper, we introduce a structural change model with Poisson process into a system that outputs a set of ongoing time series data, moment by moment. The model can be considered as a kind of Hidden Markov Model. According to the model, we formulate a method to find out the structural change, by defining a New Sequential Probability Ratio (NSPR), which can be calculated from the joint occurrence probability of the observing event with the event H0 (the structural change is not occurred) and H1 (the change is occurred). And also, we show the simple recurrence equation of the NSPR.

Copyright
© 2013, the Authors. Published by ALife Robotics Corp. Ltd.
Open Access
This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).