Sentiment Analysis of Patient Reviews of Az-Zainiyah Clinic Services Using Neural Language Processing with the Naïve Bayes Method

Abstract

In the research, the researcher evaluates analysis sentiment from review patients about services at the Az-Zainiyah clinic with the use Naïve Bayes method in Natural Language Processing (NLP). The dataset used consists of from review grouped patients become three categories of sentiment: positive, neutral, and negative. The Naïve Bayes model was trained and tested. To test its performance in classifying sentiment review patients. Research results show that the Naïve Bayes model achieves accuracy by 96%, Good macro average or weighted average shows high precision, recall, and f1-score values, around 0.97 and 0.96, respectively. These results show the effectiveness of the model in identifying sentiment review patients with high accuracy. Findings This gives valuable insights for increased quality services at the Az-Zainiyah clinic based on bait come back from patients, who in turn can increase satisfaction and experience patient.