Abstract:
In this dissertation, three new dialectometric methods for measuring the phonetic differences between the dialects of a language are introduced: an application based on vector analysis, an approach using bigrams and several algorithms from the field of information theory. These methods are applied to a Bulgarian dialect data set and the data is visualized with several different techniques. The results of these examinations are analyzed with methods from unsupervized machine learning and compared to the assertions of traditional Bulgarian dialectology. Finally, a dialect data set from northern Italy is analyzed with approved methods of dialectometry.