HYPERPARAMETER TUNING OPTIMIZATION USING GRIDSEARCH AND RANDOMSEARCH IN SUPPORT VECTOR MACHINE (SVM) METHOD FOR THE CLASSIFICATION OF HEART DISEASE PATIENT DATA

DINDA, MEILANI ADITYA WATI (2025) HYPERPARAMETER TUNING OPTIMIZATION USING GRIDSEARCH AND RANDOMSEARCH IN SUPPORT VECTOR MACHINE (SVM) METHOD FOR THE CLASSIFICATION OF HEART DISEASE PATIENT DATA. FAKULTAS MATEMATIKA DAN ILMU PENGETAHUAN ALAM, UNIVERSITAS LAMPUNG.

[img]
Preview
 Text
1. ABSTRAK - ABSTRACT.pdf
Download (10kB) | Preview
[img] Text
2. SKRIPSI FULL.pdf
Restricted to Repository staff only
Download (2MB) | Request a copy
[img]
Preview
 Text
3. SKRIPSI TANPA BAB PEMBAHASAN.pdf
Download (2MB) | Preview

Abstract

Heart disease is one of the most deadly Non-Communicable Diseases (NCD) in the world and is the leading cause of death in Indonesia. Efforts to early detection of heart disease are very important to increase the effectiveness of treatment and reduce the risk of death. In this study, the Support Vector Machine (SVM) method is used to classify data of heart disease patients. This study aims to optimize the performance of SVM models through hyperparameter tuning techniques using GridSearch and RandomSearch, as well as the application of data balancing methods, namely Random Oversampling. The data used is secondary data obtained from the Kaggle site, consisting of 918 patient data with 11 independent variables and 1 target variable. The research process includes data preprocessing stages (cleaning, feature selection, balancing, scaling), modeling with SVM, and model performance evaluation using accuracy, precision, recall, and f1-score metrics. The results show that hyperparameter tuning optimization using RandomSearch with Random Oversampling technique produces the best performance with an accuracy of 89.21%, compared to GridSearch and the default model which has an accuracy of 85.29%. Thus, RandomSearch is proven to be more effective in improving the performance of SVM classification models to detect heart disease. Keywords: Support Vector Machine, Hyperparameter Tuning, GridSearch, RandomSearch, Heart Disease, Random Oversampling.

Item Type: Other
Subjects: ?? 510 ??
Divisions: ?? matematika ??
Depositing User: 2507577324 Digilib
Date Deposited: 16 Oct 2025 02:09
Last Modified: 16 Oct 2025 02:09
URI: http://digilib.unila.ac.id/id/eprint/91225

Actions (login required)

View Item View Item