ANALISIS PERBANDINGAN ALGORITMA C4.5 DAN NAIVE BAYES UNTUK MEMPREDIKSI KETERCAPAIAN TARGET PO DALAM MEMBANGUN PROJECT FTTH (FIBER TO THE HOME)

In the digital era, the demand for high-speed and stable internet has become essential to support communication and information access. Fiber to the Home (FTTH) is one of the main solutions implemented by internet service providers such as MyRepublic. A critical component in FTTH network development is the issuance of Purchase Orders (PO) to vendors, which directly impacts the achievement of sales targets. This study aims to compare the performance of the C4.5 and Naïve Bayes classification algorithms in predicting PO target achievement to assist project planning and decision-making. The research uses historical data from FTTH projects and applies data partitioning scenarios of 70:30, 80:20, and 90:10 for model training and testing. Evaluation was conducted using accuracy as the main performance metric. The results show that the Naïve Bayes algorithm achieved the highest accuracy of 85.64% with a 70:30 data split, while C4.5 obtained 83.54% accuracy with a 90:10 data split. Based on these findings, the Naïve Bayes algorithm is considered more effective and consistent in predicting PO target achievement and is recommended for implementation in similar project scenarios.