Developing Data Mining Prediction System for Health Center Medicine Inventory using Naïve Bayes Classifier Algorithm

Abstract

Public health centers mostly use conventional methods in managing drug supply, usage, and demand data, without a system that can predict the number of drug requests. This research aims to develop a data mining solution by implementing a prediction system using the Naïve Bayes Classifier algorithm to predict drug supplies from the Koni Health Center, Jambi, to the Health Office Pharmacy Installation. The method applied in this research is a quantitative approach through the experimental method. The research data includes inventory, usage, and remaining stock of various types of drugs from 2017 to 2021 which are divided into four quarters. The results of this study show that the classification system using the Naïve Bayes Classifier method is able to classify data quickly and efficiently according to drug supply. The system test results show an accuracy of 73.91%, recall of 85.71%, and precision of 54.54%. These findings can help Puskesmas in optimizing drug inventory management, reducing errors in inventory estimates, and increasing accuracy in meeting patient drug requests.