Peramalan Jumlah Kasus Baru HIV Menurut Provinsi Menggunakan Machine Learning dengan Teknik Levenberg-Marquardt

Abstract

Early detection of HIV is a crucial step to reducing transmission and increasing the success of HIV treatment. The sooner HIV is detected, the sooner treatment can be carried out so that this infection can be controlled and does not develop into AIDS. Therefore, the purpose of this study is to forecast the Number of New HIV Cases in Indonesia based on 34 Provinces so that the government can obtain information early on to determine the right policy to suppress the increasing number of new HIV cases in Indonesia. This research proposes forecasting using a Machine Learning algorithm with the Levenberg-Marquardt technique. The research data is data on the number of new HIV cases by province obtained from the 2021 Indonesian Health Profile book issued by the Ministry of Health of the Republic of Indonesia. This research will be analyzed using three network architecture models, 3-15-1, 3-20-1 and 3-25-1. Based on the analysis of the three models used, the results show that the 3-15-1 model is the best because it produces a higher accuracy level than the other two models, which is 88%. It can be concluded that the Levenberg-Marquardt technique with the 3-15-1 model is quite suitable for forecasting new cases of HIV in Indonesia. Based on the prediction results, the number of new HIV cases by the province in Indonesia at the end of 2022 decreased significantly compared to 2021, which was 24668 compared to 36902 or reduced by around 12 thousand cases.