Detecting changes in linear regression models with skew normal errors

Abstract

Abstract In this article, we discuss a linear regression change-point model with skew normal errors. We propose a testing procedure, based on a modified version of the Schwarz information criterion, which is named the modified information criterion (MIC) to locate change points in such a linear regression model. Due to the difficulty of derivation of the asymptotic null distribution of the associated test statistic analytically, the empirical critical values at different significance levels are approximated through simulations. Simulations have also been conducted under different changes among parameters of interest with various sample sizes to investigate the performance of the proposed test. Such a procedure has been applied on a NASA data to illustrate the detecting process.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.