Research Article

Machine Learning in Public Health: Cluster Analysis of Dengue Fever Periode 2012-2020 in Indonesia

Authors: Rachmah Indawati, Mochammad Bagus Qomaruddin.

DOI: 10.33425/2833-0382.1034

---

Abstract

Machine Learning offers a powerful tool for tracking diseases efficiently, quickly, and flexibly. This study aims to analyze sequential datasets 2012-2020 on Dengue fever, providing new insights into the latest trends and developments related to the disease. First, a review of health profile reports from East Java. Second, an examination of disease data, including the number of cases, deaths, and the Case Fatality Rate, CFR. Third, clustering analysis using the k-means method to identify regions with similar or distinct characteristics based on the number of cases, deaths, and CFR.

The results reveal that regions with low case counts are those with numbers below the East Java average. These low-case regions are more prevalent compared to those with moderate or high case counts. Similarly, clustering based on the number of deaths shows that regions with low death counts are more common. However, between 2012-2020, certain areas in East Java experienced high death counts and CFR clusters. Notably, some regencies/ cities within high CFR clusters have seen repeated occurrences. A region’s CFR cluster can change over time, shifting from low to moderate or high, and vice versa.

The findings highlight that the high number of regions with elevated death clusters underscores the serious challenges East Java faces in combating Dengue Fever. The recurrence of high CFR clusters in certain areas indicates the need for sustained vigilance against the severity of the disease. Additionally, the dynamic changes in cluster classifications across regions suggest that Dengue fever is highly contagious. The widespread distribution of the disease across nearly all regencies and cities in East Java implies potential spatial or regional factors influencing its spread.

View / Download PDF