Jurnal Penelitian Rumpun Ilmu Teknik
Vol. 2 No. 3 (2023): Agustus : Jurnal Penelitian Rumpun Ilmu Teknik

Optimasi Feature Selection Menggunakan Algoritma Neural Network Untuk Klasifikasi Brain Stroke

Serly Agustin (Universitas Adhirajasa Reswara Sanjaya)
Rizkia Meinita (Universitas Adhirajasa Reswara Sanjaya)
Fiqri Khalid Aziz Al-rasyid (Universitas Adhirajasa Reswara Sanjaya)
Amelia Anjani (Universitas Adhirajasa Reswara Sanjaya)
Rehan Alif Albani (Universitas Adhirajasa Reswara Sanjaya)
Ricky Firmansyah (Universitas Adhirajasa Reswara Sanjaya)



Article Info

Publish Date
23 Jun 2023

Abstract

One of the deadliest strokes is a brain stroke. According to the results of many cases of brain stroke patients, there is a possibility that bad lifestyles such as smoking and drinking alcohol can cause high blood pressure. The goal is to classify triggers for brain structure symptoms by comparing several algorithms. From the results of this comparison, it is possible to obtain triggers with the highest number of triggers so that later brain structures can be diagnosed more quickly. In several algorithms namely nn , feature selection and GA. To group triggers for several brain stroke symptoms, to maximize feature weight and feature selection, data processing using rapidminer was continued with four algorithms: X-Fold validation and split validation with ratios of 0.5, 0.6, 0.7, 0.8 and 0.9. After this test, the most popular AUC values and methods, together with the Neural Net algorithm, the Optimize Selection (Evolutionary) feature, and using a Split Validation ratio of 0.9, produce numbers with very high accuracy. AUC of 0.549 and an accuracy value of 95.88%..

Copyrights © 2023






Journal Info

Abbrev

JUPRIT

Publisher

Subject

Computer Science & IT Engineering

Description

Sistem Komputer / Teknik Komputer Sistem Informasi Teknik Perangkat Lunak Teknologi Informasi Teknik Informatika / Ilmu Komputer Bidang-bidang lainnya yang termasuk ke dalam rumpun ilmu ...