CLASSIFICATION OF REGIONAL LANGUAGES USING METHODS GRADIENT BOOTS AND RANDOM FOREST

Indonesia is one of the countries that has the most regional languages ​​in the world, ranking second most. The large number of regional languages ​​that are owned makes it difficult for people between regions to recognize the origins of the regional language, so the author aims to conduct research by identifying a regional language. Identifying a language using data mining, one of the data mining techniques is classification. Classification is a technique used to find the value of data. Classification will build a model from samples of data into groups of the same type. There are two classification methods used in this research, namely gradient boots and random forest, where the two methods will be compared using regional language data from Java, Nias and Toraja. The results of calculating the accuracy values ​​for the two methods used are quite good in classifying languages ​​with results of an accuracy level of 0.8 or 80%, where the results of the gradient boots research have an accuracy value of 0.8850 or 88.5%, while the random forest method has an accuracy value. random forest is lower, namely 0.8794 or 87.94%, so in this study the gradient boots method is the best method.