Riza Arifudin, Vincentius Natanael Siahaan, Abas Setiawan
Huge amounts of data are collected every day; consciously or not, humans produce petabytes of data, which are contained in various storage devices in business, social, science, industry, health, and other aspects of life. When viewed briefly, this large amount of data has no meaning, but after processing it, the data will reveal information of great value that can be used for the benefit of human life. This research uses the Wisconsin Diagnostic Breast Cancer dataset from the UCI Machine Learning Repository. The WDBC dataset has 569 rows and 32 columns, and it has a high-dimensional classification role. This research aims to find an efficient classification model by selecting several features to obtain greater accuracy. Gaussian Naive Bayes, which is a classification method, can produce a model with a cross-validation value score of 93.677%. The Wrapper Forward Selection feature selection method combined with Gaussian Naive Bayes succeeded in selecting 12 features that produced the highest Cross-Validation score, namely 97.538%. So, the increase in accuracy obtained by using forward selection and Gaussian naive Bayes is 3.861%. © 2025 Author(s).
Computer Science Department, Universitas Negeri Semarang, Semarang, Indonesia
