ARDIANSAH, VALENT (2026) PERANCANGAN SISTEM PREDIKSI RISIKO ISPA BERBASIS WEB MENGGUNAKAN RANDOM FOREST DENGAN OPTIMASI HYPERPARAMETER. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
Cover.pdf Download (443kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
Bab 1.pdf Restricted to Registered users only Download (33kB) |
|
|
Text (BAB II)
Bab 2.pdf Restricted to Registered users only Download (178kB) |
|
|
Text (BAB III)
Bab 3.pdf Restricted to Registered users only Download (139kB) |
|
|
Text (BAB IV)
Bab 4.pdf Restricted to Registered users only Download (808kB) |
|
|
Text (BAB V)
Bab 5.pdf Restricted to Registered users only Download (27kB) |
|
|
Text (DAFTAR PUSTAKA)
Daftar Pustaka.pdf Restricted to Registered users only Download (133kB) |
|
|
Text (LAMPIRAN)
Lampiran.pdf Restricted to Registered users only Download (197kB) |
Abstract
Acute Respiratory Infection (ARI) requires early detection to prevent serious clinical complications caused by limited healthcare access and low public awareness. Furthermore, the high potential for rapid disease spread within the community demands an easily accessible self-screening solution. Addressing this urgency, this study proposes a web-based ARI risk prediction system as a fast and accurate real-time independent screening tool. This study utilizes 396 secondary medical record data points consisting of risk factors and clinical symptoms. Considering the initial class imbalance issue within the dataset with a 71:29 ratio between risk and non-risk categories, this study applies the Synthetic Minority Over-sampling Technique (SMOTE) on the training data—successfully balancing the dataset to 1,138 samples per class—alongside the class_weight='balanced' parameter configuration in the Random Forest algorithm to prevent model bias. The applied classification method is multi-class classification using the Random Forest algorithm with hyperparameter optimization via Grid Search and 5-fold cross-validation. The web-based system is designed using Flask architecture with integrated UML modeling and tested through Black Box Testing. The evaluation results demonstrate that the optimized model achieves an accuracy of 95.71%, precision of 96%, recall of 96%, F1-score of 96%, and a 0% false negative rate for the positive class. These findings prove that proper handling of data imbalance makes the developed system highly effective and reliable in supporting targeted preventive decision-making. Keywords: ARI, Random Forest, Grid Search, Early Detection System, Risk Prediction Infeksi Saluran Pernapasan Akut (ISPA) memerlukan deteksi dini guna mencegah komplikasi klinis serius akibat keterbatasan akses layanan dan rendahnya kesadaran masyarakat. Selain itu, tingginya potensi penyebaran kasus yang cepat di tengah masyarakat menuntut adanya solusi skrining mandiri yang mudah diakses. Menjawab urgensi tersebut, penelitian ini mengusulkan perancangan sistem prediksi risiko ISPA berbasis web sebagai alat skrining mandiri yang cepat dan akurat secara real-time. Penelitian ini menggunakan 396 data sekunder rekam medis dengan variabel faktor risiko dan gejala klinis. Mengingat distribusi data awal yang tidak seimbang (class imbalance) dengan rasio 71:29 antara kelas berisiko dan tidak berisiko, penelitian ini menerapkan teknik Synthetic Minority Over-sampling Technique (SMOTE) pada data latih yang berhasil menyeimbangkan jumlah data menjadi 1.138 sampel untuk masing-masing kelas, serta pengaturan parameter class_weight='balanced' pada algoritma Random Forest guna mencegah bias model. Metode klasifikasi multi-kelas yang diterapkan menggunakan algoritma Random Forest dengan optimasi hyperparameter Grid Search dan 5-fold cross-validation. Sistem berbasis web dirancang melalui arsitektur Flask dan pemodelan UML terintegrasi, serta diuji menggunakan Black Box Testing. Hasil pengujian menunjukkan bahwa model teroptimasi mencapai tingkat akurasi 95,71%, precision 96%, recall 96%, F1-score 96%, serta false negative 0% untuk kelas positif. Hal ini membuktikan bahwa penanganan data imbalance secara tepat membuat sistem yang dikembangkan sangat efektif dan andal untuk mendukung pengambilan keputusan preventif secara tepat sasaran. Kata Kunci: ISPA, Random Forest, Grid Search, Sistem Deteksi Dini, Prediksi Risi
Actions (login required)
 |
View Item |
