時間序列資料變動點估計方法的探討

A Study of Change Points Analysis for Time Series Data

蔡念庭(Tsai Nien-Ting)

指導教授 ： 蔡碧紋

國立臺灣師範大學/理學院/數學系/碩士(2022年)

https://doi.org/10.6345/NTNU202201107

摘要

當序列發生統計特性變化時，則會存在變動點。變動點檢測可用於估計序列中單個或多個變動點的位置與其資料的統計特性。本文討論在時間序列AR(1)資料下，使用Pruned Exact Linear Time(PELT)及結構變化模型(structural change model)方法找變動點。以模擬方式比較兩種不同方法在單個及多個變動點情況下，變動點檢測的結果及在不同評估準則的優劣，並且將兩種方法應用於美國COVID-19實際資料。

關鍵字

變動點檢測 ； 結構變化模型 ； Pruned Exact Linear Time

並列摘要

Changepoints occur at where statistical properties of the data change. Changepoint detection is able to estimate single or multiple changepoints in the series. In this thesis, we consider the change point detection problem for AR(1) time series data. Pruned Exact Linear Time(PELT)and structural change model methods are used to find the location of the change points. Some simulation studies are done and several criteria are used to compare the results. Additionally, an application to the United States COVID-19 data is presented.

並列關鍵字

Changepoint Detection ； Pruned Exact Linear Time ； Structural Change Model

參考文獻

Akaike, H. (1974). A new look at the statistical model identification. IEEE transactions on automatic control, 19(6), 716-723.

Auger, I. E., & Lawrence, C. E. (1989). Algorithms for the optimal identification of segment neighborhoods. Bulletin of mathematical biology, 51(1), 39-54.

Bai, J. (1994). Least squares estimation of a shift in linear processes. Journal of Time Series Analysis, 15(5), 453-472.

Bai, J., & Perron, P. (1998). Estimating and testing linear models with multiple structural changes. Econometrica, 66(1), 47-78.

Bai, J., & Perron, P. (2003). Computation and analysis of multiple structural change models. Journal of applied econometrics, 18(1), 1-22.