Discovery of association rules between syntactic variables

Abstract

This research applies an association rule mining technique to purely syntactic dialect data. The paper answers the research question of how relevant associations between syntactic variables can be discovered. The method calculates the proportional overlap between geographical distributions of syntactic microvariables and incorporates rule quality factors such as accuracy, coverage and completeness to measure the interestingness of the variable associations. The exploratory review of the results discusses several highly ranked association rules and also examines an implicational chain of syntactic variables.