CHOI Woonho. (2010). A Computational Approach to Dialectometry and its Application to the Mongolic Languages and Dialects in China. Altai Hakpo 20: 127-160. The Altaic Society of Korea.

In recent times, many computational approaches to dialectology have been developed. Dialectometry is a study on measuring the distances between dialects and languages, and on the classification and clustering of those languages and dialects. In this study, 16 Mongolic dialects and languages in China were analyzed and compared to each other by pairwise comparison methods. In comparison, the Levenshtein distance method is used to measure the distance between two words, and its value is normalized by the length of aligned words. After measuring the distance between among languages and dialects, the distance matrix is created. The distance matrix is analyzed by the multidimensional scaling, the hierarchical clustering methods, and some phylogenetic analysis methods. The clustering result says that some dialects are closer to another dialects, and others are not. Especially, it is shown that BULIYATE and CHENBARHU is the closest dialects.

Keywords: omputational linguistics, dialectometry, Levenshtein distance, hierarchical clustering, multidimensional scaling, Mongolic languages, dynamic programming