The Minimum Description Length Guided Model Selection in Granger Causality Analysis

In data analysis and statistical modeling, the optimal model selected is the key to success. This article is mainly talked about the minimum description length applied in linear models to select and optimize models, which the minimum description principle could be also applied in many other model classes at the same time. The principle would give a more suitable methods in modeling and estimating. Combining with Granger causality, it's meanful to understand brain activities deeply.