Analisis Performa Algoritma NBC, DT, SVM dalam Klasifikasi Data Ulasan Pengunjung Candi Borobudur Berbasis CRISP-DM

The approach of visitor sentiment analysis to Borobudur Temple tourist destinations in Indonesia can be classified using various algorithms get optimal results. Good algorithm performance seen from the confusion matrix (accuracy, precision, recall) value, Area Under Curve (AUC) and Receiver Operating Characteristic (ROC). This study used Naïve Bayes Classifier (NBC), Decision Tree (DT), Support Vector Machine (SVM) against 3850 text data obtained Tripadvisor website, especially reviews visitors. method refers Cross-Industry Standard Process for Data Mining (CRISP-DM) optimizing destination products services by paying attention six stages: business understanding, preparation, modeling, evaluation, deployment. results this show that NBC's evaluation have a change value at accuracy 98.73% 95.6%, precision changed 98.72% 98.97%, recall also 100% 96.54%. In addition, NBC 0.500 (50%) 0.693 (69.35%). DT showed 97.55% 94.40%, increased 97.63% 91.86%, 99.90% 99.47%. 0.591 (59.1%) 0.932 (93.2%). SVM 99.41%; 100%, 99.01%. 0.961 (96.1%) 1.00 (100%). T-test is more dominant compared other algorithms, where 0.994 0.944 0.98. Based on (ROC) it shows good addition SVM. It indicates analyzing visitors Temple, best-recommended