Particle Swarm Optimization For Improved Accuracy of Disease Diagnosis

Suamanda Ika Novichasari (Ngudi Waluyo University)
Iwan Setiawan Wibisono (Ngudi Waluyo University)

JAIS (Journal of Applied Intelligent System)
Vol 5, No 2 (2020): Journal of Applied Intelligent System

The increasing number of patients suffering from various diseases and the impact and high cost of medical treatment for the community has made the government or health communities seek solutions for prevention from an early age. This valuable information can be found using artificial intelligence and data mining. Most diseases are dangerous; if detected early and adequate diagnosis and treatment are available, there will be a chance for a cure. The main objective of this study was to use Particle Swarm Optimization (PSO) to improve the accuracy of several classification methods, namely Naive Bayes, C4.5, Support Vector Machine (SVM), and Neural Network (NN) to detect heart disease, hepatitis, kidney, and breast cancer. The method used in this research is the CRISP-DM model, with 5 stages. The data used were four disease data from UCI Machine Learning. The result of this research is that PSO can improve the accuracy of Naive Bayes, C4.5, SVM, and NN.