MOBILITY ANALYSIS AND PREDICTION OF CUSTOMER DISTRIBUTION FOR BPJS KESEHATAN USING THE MARKOV CHAIN METHOD

• Main Author: Pratisena, Andhika
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84741
• Institution: Institut Teknologi Bandung

BPJS Kesehatan is the provider of social health insurance in Indonesia. With a different system from general insurance companies, BPJS Kesehatan covered 95.75% of the Indonesian population in 2023. Despite this, BPJS Kesehatan is still obliged to maintain its company’s financial stability, one way being through the control of revenue. The revenue received by BPJS Kesehatan is no longer relevant if predicted using time series because the addition of revenue is not significantly influenced by the increase in customers. Therefore, an analysis was conducted on BPJS Kesehatan participants based on customer segmentation that indirectly affects the amount of revenue. Using participant mutation data from 2017-2022, this study predicts participant distribution based on categories using the Markov chain method applied to an automation program in Python language. In line with the prediction calculations, a transition matrix was obtained which describes the probability of movement within customers’ categories. Through this matrix, dynamic analysis can be performed to generate useful information for the company's operational strategy. Prediction results show that the Markov chain method has the highest accuracy when using the most recent year's dataset as the prediction baseline. Predictions in this study show a good accuracy level with an RMSE of 0.941% for modeling participant distribution.