CLASSIFYING VILLAGE FUND IN WEST JAVA, INDONESIA USING CATBOOST ALGORITHM

Abstract

With over 261 million inhabitants, Indonesia is home to approximately 15,000 villages, according to the Ministry of Villages, Disadvantaged Regions, and Transmigration. Among these, 1,406 are in West Java. Of these, 504 of them are advanced, 464 are developing, 390 are disadvantaged, and 48 are very disadvantaged. The CatBoost machine learning model was used to classify village funds in West Java from 2018 to 2021 and had an accuracy rating of 75%, precision rating of 79%, recall of 79%, and f1 score of 79%, demonstrating its excellent performance. However, missing data points had to be removed from the analysis and it is suggested that a more sophisticated method for handling missing values should be used in future studies. In addition, hyperparameter tuning could be employed to increase the model's performance, and a variety of metrics could be used to accurately assess the results. Overall, CatBoost may be of benefit to the Indonesian Government in order to classify village funds according to their status, channel funds more accurately and efficiently, and observe the situation of a village year-over-year.