Enhancing Accuracy by Using Boosting and Stacking Techniques on the Random Forest Algorithm on Data from Social Media X

Abstract

Online loans (commonly referred to as Pinjol) have become a widespread phenomenon in Indonesia, both in legal and illegal forms. It is undeniable that this is in line with the rapid development and innovation of technology. Pinjol cannot be separated from public comments, both positive and negative, on social media X. The study examined the communication patterns of Indonesian people using a sentiment analysis approach. The research utilized the Random Forest algorithm to perform sentient analysis. This algorithm combined the output of several decision trees to achieve a more accurate result. In addition to using a random forest algorithm, this study also made improvements by using stacking and boosting. The results of this study indicated that the highest accuracy of 86% was obtained by the SMOTE+RF+Adaboost (Boosting) model. In contrast, the lowest accuracy of 60% was obtained in the RF+Adaboost model with a stacking technique.

Keywords

• boosting
• random forest
• smote
• social media x
• stacking