Pengembangan Sistem Prediksi Diagnosis Penyakit Jantung Dan Sistem Pakar untuk Pasien Gagal Jantung

Sunandar, Naufal Sirrulah Ahmad (2023) Pengembangan Sistem Prediksi Diagnosis Penyakit Jantung Dan Sistem Pakar untuk Pasien Gagal Jantung. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 05211940000104-Undergraduate_Thesis.pdf] Text
05211940000104-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2025.

Download (10MB) | Request a copy

Abstract

Penyakit Kardiovaskular atau penyakit jantung adalah penyakit yang sangat fatal. Menurut data World Health Organization (WHO) dan Kementerian Kesehatan Indonesia (Kemenkes), penyakit jantung dikategorikan sebagai penyebab kematian tertinggi di dunia. Dalam upaya mitigasi penyakit jantung, perkembangan teknologi telah memberikan alat yang dapat membantu dalam diagnosis dini penyakit jantung dan pemilihan terapi yang tepat untuk mencegah gagal jantung. Penelitian ini bertujuan untuk membangun sistem prediksi penyakit jantung dan sistem pakar untuk pasien dengan gagal jantung. Sistem prediksi penyakit jantung dikembangkan menggunakan teknik machine learning dengan menggunakan algoritma klasifikasi Support Vector Machine (SVM). SVM dipilih sebagai model prediksi karena SVM menunjukkan tingkat akurasi yang tinggi dan telah banyak diimplementasikan dalam berbagai aplikasi, seperti klasifikasi email spam, pengenalan wajah, dan pengenalan gen. Dataset yang digunakan untuk membangun sistem prediksi ini adalah “HEART DISEASE DATASET (COMPREHENSIVE)” yang dikumpulkan oleh Manu Siddhartha, seorang mahasiswa program Master of Science dalam Artificial Intelligence and Machine Learning di Liverpool John Moores University. Sementara itu, sistem pakar untuk penyakit gagal jantung dibangun menggunakan prolog dengan knowledge base utama yang berasal dari jurnal "2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines". Jurnal ini merupakan rekomendasi yang diberikan oleh dokter mitra penelitian yang merupakan spesialis jantung dan pembuluh darah (S. JP) yang bekerja di salah satu rumah sakit di Surabaya. Dalam tahap pengembangan sistem prediksi, dataset melalui proses pembersihan, implementasi teknik SMOTE, scaling, dan seleksi fitur. Pada akhirnya, model sistem prediksi mencapai akurasi sebesar 92.25%. Sementara itu, dalam pembangunan sistem pakar, rekomendasi dari jurnal diubah menjadi format data CSV dengan total 102 rekomendasi. Data ini kemudian dikonversi menjadi sistem prolog yang terdiri dari aturan (rules) dan fakta (facts). Pada tahap akhir pengembangan sistem pakar, terdapat total 136 aturan dengan 4 parameter fakta yang mencakup bukti dan diagnosis (evidence and diagnosis), riwayat (history), pengukuran (measurement), dan intoleransi (intolerant) dari pasien. Sistem pakar juga dapat menampilkan rekomendasi berdasarkan fase pasien dengan mengikuti prioritas COR (Class of Recommendation) dan LOE (Level of Evidence). Hasil akhir dari penelitian ini adalah implementasi sistem prediksi dan sistem pakar dalam bentuk aplikasi web menggunakan framework Flask. Diakhir tahap kami melakukan survei pengguna untuk kedua sistem, dan secara umum tingkat kepuasan pengguna terhadap sistem pakar mencapai 6.5/10, sedangkan kepuasan terhadap sistem prediksi mencapai 8.5/10
=====================================================================================================================================
Cardiovascular disease or heart disease is a highly fatal condition. According to data from the World Health Organization (WHO) and the Ministry of Health of Indonesia (Kemenkes), heart disease is categorized as the leading cause of death worldwide. In an effort to mitigate heart disease, technological advancements have provided tools that can assist in the early diagnosis of heart disease and the selection of appropriate therapies to prevent heart failure. This research aims to develop a predictive system for heart disease and an expert system for patients with heart failure. The predictive system for heart disease is developed using machine learning techniques with the Support Vector Machine (SVM) classification algorithm. SVM is chosen as the predictive model due to its high accuracy and widespread implementation in various applications, such as spam email classification, face recognition, and gene recognition. The dataset used to build this predictive system is the "HEART DISEASE DATASET (COMPREHENSIVE)" collected by Manu Siddhartha, a Master of Science student in Artificial Intelligence and Machine Learning at Liverpool John Moores University. On the other hand, the expert system for heart failure is built using Prolog with the main knowledge base derived from the journal "2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines." This journal is a recommendation provided by the research partner, who is a cardiovascular specialist (S. JP) working at a hospital in Surabaya. During the development phase of the predictive system, the dataset undergoes cleaning, the implementation of SMOTE technique, scaling, and feature selection. Ultimately, the predictive system model achieves an accuracy of 92.25%. In the development of the expert system, the recommendations from the journal are converted into CSV data format with a total of 102 recommendations. This data is then converted into a Prolog system consisting of rules and facts. In the final stage of expert system development, there are a total of 136 rules with 4 fact parameters that cover evidence and diagnosis, history, measurements, and intolerances of the patient. The expert system can also display recommendations based on the patient's phase, following the COR (Class of Recommendation) and LOE (Level of Evidence) priorities. The end result of this research is the implementation of the predictive system and expert system in the form of a web application using the Flask framework. At the end of the process, a user survey was conducted for both systems, and overall, the user satisfaction level with the expert system reached 6.5/10, while satisfaction with the predictive system reached 8.5/10.

Item Type: Thesis (Other)
Uncontrolled Keywords: Penyakit Jantung, SVM, Sistem Prediksi, Sistem Pakar, Gagal Jantung, Heart Disease, SVM, Prediction System, Expert System, Heart Failure
Subjects: Q Science > QA Mathematics > QA336 Artificial Intelligence
Q Science > QA Mathematics > QA76.76.E95 Expert systems
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User: Naufal Sirrullah Ahmad Sunandar
Date Deposited: 01 Aug 2023 00:34
Last Modified: 01 Aug 2023 00:34
URI: http://repository.its.ac.id/id/eprint/101008

Actions (login required)

View Item View Item