PENGUJIAN ALGORITMA TEKS MINING UNTUK KLASIFIKASI ANALISIS REVIEW APLIKASI HALODOC

Abstract

Entering the era of the industrial revolution 4.0 technology is developing very rapidly, information system technology in the health sector is e-health as information and communication technology that is effective and safe in support matters related to the health sector such as health services, health supervision, references on matters relating to health. health matters. Consumers who write reviews, opinions, and experiences in medical teleconsultation continue to increase. Halodoc dataset must be processed using the right algorithm. So the results of this study are to find out which algorithm is better used to get the best algorithm model. Researchers compared several Mining Text classifications, including the C4.5 Algorithm, K-Nearest Neighbor (K-NN), and Support Vector Machine (SVM). The stages of the research carried out were starting from data collection, initial data processing (C4.5, K-NN, and SVM methods), the testing method used was 10-fold cross-validation, evaluation results, and testing using t-test, the proposed method. From the process that has been carried out, the accuracy results obtained using the K-Nearest Neighbor (K-NN) algorithm are 88.50% with an AUC value: of 0.960. Meanwhile, the best model results use the t-test test, namely the algorithm: Support Vector Machine algorithm and the K-NN algorithm in testing the Halodoc dataset.