ANALISIS PENGARUH PENGURANGAN DIMENSI DATA PADA KEAKURATAN PREDIKSI PENYAKIT JANTUNG DENGAN MENGGUNAKAN SVM LINEAR

Akhmad Ghiffary Budianto (Universitas Lambung Mangkurat)
Akhmad Syarief (Universitas Lambung Mangkurat)

Jurnal Taguchi : Jurnal Ilmiah Teknik dan Manajemen Industri
Vol. 3 No. 1 (2023): Jurnal Taguchi : Jurnal Ilmiah Keilmuan Teknik dan Manajemen Industri

Heart disease is a disorder in the form of plaque that occurs in large blood vessels. This disrupts the supply of oxygen to the organs of the body. Heart disease is 1 of the 3 most common causes of death worldwide. Therefore, early detection based on the examination of medical data is needed to prevent the impact. The method used for classification is Support vector machine (SVM) and dimension reduction is Principal component analysis (PCA). The dataset is from Kaggle, medical records of 299 patients with 12 features and 1 label. The results obtained are the level of accuracy of PCA 6 features and without PCA both produce 82.9% and a total of 51 misclassifications. The processing time required is slightly longer for PCA 6 features (0.69121 seconds) than without PCA (0.46173 seconds). Because it has the same level of accuracy, the f-score metric is used to assess the classification model. The SVM with PCA 6 features has an f-score of 0.879, this is slightly better than SVM without PCA, which is 0.878