Prospect Prediction Model of Indonesian Telematics Medium and Large Enterprises Using Ensemble Approach

Abstract

Medium and Large Enterprises (MLE) play an important role in promoting development and increasing Indonesia's competitiveness in the international arena. Currently, there is not much in-depth analysis related to the prospects for medium and large telematics businesses to find out its relevance to economic conditions and what factors affect business prospects. Whereas the telematics industry sector is a supporting sector for Gross Domestic Product (GDP) with a higher growth rate compared to other sectors. The purpose of this study is to predict the prospects MLE telematics businesses using an ensemble approach with the Boost and Bagging algorithm. Data obtained from the 2016 Economic Census. The target class for business prospects consists of five classes, namely better, the same as bad, worse, and cannot be compared. Optimal results are obtained with the average precision value is 76,8, recall is 79,0 and F1-score is 77,13. The model is able to produce important features that can be validated by experts. The advantages of the prospective MLE telematics feature are the level of sales of goods and services, profits, and the level of education of the business owner.