Pemanfaatan Machine Learning untuk Peningkatan Akurasi Sistem Pendukung Keputusan Prediktif

Abstract

The rapid development of information technology and the increasing availability of large-scale data have driven the need for decision-making systems that are more intelligent, faster, and more accurate. Conventional Decision Support Systems (DSS) generally rely on rule-based approaches or simple statistical analyses, which have limitations in recognizing complex patterns and are less adaptive to changes in data. Therefore, the integration of machine learning technology represents a strategic solution to enhance the predictive capability and decision quality produced by DSS. This study aims to analyze the utilization of machine learning algorithms in improving the accuracy of predictive decision support systems. The method employed is a comparative experimental approach involving three algorithms, namely Decision Tree, Random Forest, and Support Vector Machine. The dataset used consists of historical decision outcomes along with their determining variables derived from a case study. The research stages include data cleaning, normalization, training–testing set splitting, and evaluation using accuracy, precision, recall, and F1-score metrics. The results indicate that the application of machine learning significantly improves DSS accuracy compared to conventional methods. Random Forest achieved the best performance with an accuracy of 91%, followed by Support Vector Machine at 87% and Decision Tree at 84%. In addition to improving accuracy, the integration of machine learning also enhances data processing efficiency and decision-making speed. These findings demonstrate that artificial intelligence–based DSS has great potential for application across various domains, such as business, healthcare, education, and government.