Data mining using a support vector machine, decision tree, logistic regression and random forest for pneumonia prediction and classification

Bahtiar Imran (Universitas Teknologi Mataram, Mataram and 83115, Indonesia)
Zaeniah (Universitas Teknologi Mataram, Mataram and 83115, Indonesia)
Sriasih Sriasih (Universitas Teknologi Mataram, Mataram and 83115, Indonesia)
Surni Erniwati (Universitas Teknologi Mataram, Mataram and 83115, Indonesia)
Salman Salman (Universitas Teknologi Mataram, Mataram and 83115, Indonesia)

INFOKUM
Vol. 10 No. 02 (2022): Juni, Data Mining, Image Processing, and artificial intelligence

This study uses Data Mining with four classification models. The object of this research is pneumonia data. The proposed models are Support Vector Machine (SVM), Decision Tree, Logistic Regression and Random Forest. Tests have been carried out using Cross-Validation Sampling and Stratified Sampling using several Folds of 3, 10 and 20. The results obtained are Logistic Regression models get the highest and most consistent accuracy results compared to SVM, Decision Tree and Random Forest. The tests evidence this carried out with the results of Number of Folds 3 getting the AUC value of 0.990, Accuracy 0.962, F1 0.962, Precision 0.962 and Recall 0.962. Number of Folds 10 gets the AUC value of 0.991, Accuracy 0.961, F1 0.961, Precision 0.961 and Recall 0.961. Number of Folds 20 gets AUC 0.991, Accuracy 0.965, F1 0.965, Precision 0.965 and Recall 0.965. From this study, Logistic Regression got good results for predicting and classifying pneumonia.