Optimasi PSO untuk Meningkatkan Performa Algoritma C4.5 dalam Memprediksi Risiko Kesehatan Kehamilan

Nowadays, many diseases are feared to threaten the health of the body in pregnant women and their fetuses, so there is a need for early prevention. One's knowledge in predicting pregnancy health risks is important for prevention. However, precise prediction of such risks can be challenging, given the involvement of considerable and complex data. The C4.5 algorithm is widely used in data analysis. Unfortunately, its performance sometimes does not produce good accuracy. So, it needs to be improved by optimizing its structure parameters. For this reason, the purpose of this research involves the use of PSO (Particle Swarm Optimization) to find optimal parameters for the C4.5 algorithm that can increase the accuracy of pregnancy health risk prediction. The results show that the C4.5 algorithm model that has been optimized with PSO has an accuracy rate of 71.65%, and the standard C4.5 algorithm model only achieves an accuracy rate of 67.49%. There is a difference of 4.16%, which shows the superiority of the PSO optimization approach in improving prediction accuracy in the C4.5 algorithm. So, the application of the C4.5 algorithm optimized with PSO can be used as a positive implication in improving the health care of pregnant women and making more accurate medical decisions. Ultimately, this research illustrates the potential of PSO in optimizing the C4.5 data classification algorithm for health knowledge.