Creating dialect sub-corpora by clustering: a case in Japanese for an adaptive method

Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)

Abstract

We propose a pipeline through which to derive clusters of dialects, given a mixed corpus composed of di erent dialects, when their standard counterpart is su ciently resourced. The test case is Japanese, where the written standard language is su ciently equipped with adequate resources. Our method starts by detecting non-standard contents rst, and then clusters what is deemed dialectal. We report the results on the clustering of mixed Twitter corpus into four dialects (Kansai, Tohoku, Chugoku and Kyushu).