Optimalisasi Klasifikasi Kanker Payudara Menggunakan Forward Selection pada Naive Bayes

Abstract

Breast cancer is a type of malignant tumor which is still the number one killer where the process of spread or metastasis takes a long time. The number of breast cancer sufferers increases every year so that if detected or caught early, prevention can be done early so as to reduce the number of breast cancer sufferers. To reduce the risk of increasing the number of cancer patients, it is necessary to do early detection, several methods can be used to assist the early detection process such as cancer screening, or computational methods. Several machine learning methods that have been chosen to solve cases of breast cancer prediction, especially the classification algorithm, including Naive Bayes have the advantage of being simple but having high accuracy even though they use little data. Weaknesses in Naive Bayes, namely the prediction of the probability result is not running optimally and the lack of selection of relevant features to the classification so that the accuracy is low. This research is intended to build a classification system for detecting breast cancer using the Naive Bayes method, by adding a forward selection method for feature selection from the many features that exist in breast cancer data, because not all features are features that can be used in the classification process. The result of combining the Naive Bayes method and the forward selection method in feature selection can increase the accuracy value of 96.49% detection of breast cancer patients.Â

Pengaruh Komposisi Split data Terhadap Performa Klasifikasi Penyakit Kanker Payudara Menggunakan Algoritma Machine Learning

“…Algoritma Naïve Bayes menggabungkan perhitungan probabilitas [26]. Pada perhitungan menggunakan Naïve Bayes, Dataset juga dibagi menjadi data training dan testing, kemudian dicari nilai rata-rata data pelatihan.…”

Data Mining Untuk Classification Dalam Mendeteksi Kerusakan Elektrokardiogram Menggunakan Naive Bayes

“…lektrokardiogram (EKG) adalah tes untuk mengukur dan merekam aktivitas dari jantung menggunakan mesin pendeteksi impuls listrik (elektrokardiogram) [1]. Alat ini menerjemahkan impuls listrik menjadi grafik yang ditampilkan pada layar pemantau.…”

Feature Selection Menggunakan Binary Wheal Optimizaton Algorithm (BWOA) pada Klasifikasi Penyakit Diabetes

“…Salah satu langkah penting untuk menghasilan akurasi klasifikasi yang tinggi, melalui pemilihan fitur relevan yang digunakan dalam penelitian. Masalah dalam seleksi fitur adalah pengurangan dimensi, dimana awalnya semua atribut diperlukan untuk memperoleh akurasi yang maksimal sementara tidak semua fitur digunakan dalam proses klasifikasi (Astuti et al, 2020). Feature selection pada penerapannya dibagi menjadi dua kelompok ranking selection dan subset selection.…”