Authors:
Titik Misriati
1
;
Riska Aryanti
2
and
Asriyani Sagiyanto
3
Affiliations:
1
Sistem Informasi Akuntansi Kampus Kabupaten Karawang, Universitas Bina Sarana Informatika, Jakarta, Indonesia
;
2
Ilmu Komputer, Universitas Bina Sarana Informatika, Jakarta, Indonesia
;
3
Hubungan Masyarakat, Universitas Bina Sarana Informatika, Jakarta, Indonesia
Keyword(s):
Heart Disease, Classification, Support Vector Machine.
Abstract:
Heart disease is a prominent cause of mortality in developed and developing countries, including Indonesia. Conventional methods of diagnosing cardiac disease may not always be accurate, and there is an increasing demand for more modern and dependable procedures. The study aims to assess the effectiveness of several machine learning algorithms in heart disease categorization to determine the best effective algorithm for accurate diagnosis. Data mining techniques are one method for making predictions. This study employs decision tree algorithms, random forests, support vector machines, neural networks, and naive bayes to predict cardiac disease. Based on the results of the test shows that the accuracy of the Support Vector Machine algorithm is 81.97%, and the AUC 0.903 obtains higher accuracy than the Naı̈ve Bayes, Random Forest, Neural Network, and Decision Tree algorithms. Testing the Support Vector Machine algorithm using parameter C with values of 0.0, 1.0, 2.0, 3.0, 4.0, and 5.0
produces the best C parameter of 3.0 with an accuracy value of 85.25%. The results of this study, the Support Vector Machine algorithm, can be used for heart disease prediction because it has a high accuracy level and is included in the excellent classification in predicting heart disease.
(More)