Information and posterior probability criteria for model selection in local likelihood estimation

• 成果类型: Article
• 署名作者: Irizarry, RA
• 署名单位: Johns Hopkins University
• 刊物名称: JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
• ISSN/ISSBN: 0162-1459
• DOI: 10.1198/016214501750332875
• 发表日期: 2001
• 页码: 303-315
• 关键词: time-series; regression

摘要

Local likelihood estimation has proven to be an effective method for obtaining estimates of parameters that vary with a covariate. To obtain useful estimates of such parameters, approximating models are used. In such cases it is useful to consider window based estimates. We may need to choose between competing approximating models. In this article, we propose a modification to the methods used to motivate many information and posterior probability criteria for the weighted likelihood case. We derive weighted versions for two of the most widely known criteria, namely the AIC and BIG. Via a simple modification, the criteria are also made useful for window span selection. The usefulness of the weighted version of these criteria is demonstrated through a simulation study and an application to three datasets.