Muhammad Raihansyah, 26121010015 (2025) PREDIKSI KUNJUNGAN PASIEN BERDASARKAN DIAGNOSA PENYAKIT DI RUMAH SAKIT DR. BRATANATA JAMBI MENGGUNAKAN ALGORITMA DATA MINING. Sarjana thesis, Sekolah Tinggi Ilmu Kesehatan Garuda Putih.
1 COVER.docx
Download (448kB)
2 LEMBAR PERSETUJUAN.docx
Download (1MB)
3 LEMBAR PENGESAHAN.docx
Download (854kB)
4 DAFTAR RIWAYAT HIDUP.docx
Restricted to Repository staff only
Download (269kB)
5 PERNYATAAN KEASLIAN TULISAN.docx
Download (1MB)
6 ABSTRAK.docx
Download (182kB)
7 ABSTRACT.docx
Download (182kB)
8 KATA PENGANTAR.docx
Download (1MB)
9 DAFTAR ISI.docx
Download (187kB)
10 BAB I.docx
Restricted to Repository staff only
Download (201kB)
11 BAB II.docx
Restricted to Repository staff only
Download (474kB)
12 BAB III.docx
Restricted to Repository staff only
Download (197kB)
13 BAB IV.docx
Restricted to Repository staff only
Download (4MB)
14 BAB V.docx
Restricted to Repository staff only
Download (183kB)
15 DAFTAR PUSTAKA.docx
Download (193kB)
16 SURAT IZIN PENGAMBILAN DATA.docx
Download (991kB)
17 SURAT SELESAI PENELITIAN.docx
Download (708kB)
18 LEMBAR BIMBINGAN.docx
Download (1MB)
19 DATASET.docx
Restricted to Repository staff only
Download (417kB)
Abstract
ABSTRAK
Demi peningkatan kualitas layanan, pengoptimalan sumber daya, perlunya strategi manajemen pelayanan kesehatan yang berbasis data. Mengingat fluktuasi kenaikan atau penurunan kunjungan pasien yang kerap tidak terprediksi secara akurat, namun untuk mengetahui angka kunjungan pasien yang akan datang secara akurat menjadi tantangan instansi kesehatan, khusunya di Rumah Sakit. Penelitian ini bertujuan untuk memprediksi jumlah kunjungan pasien berdasarkan diagnosa penyakit di rumah sakit dr. Bratanata Jambi menggunakan pendekatan data mining dengan algoritma neural network yang diimplementasikan melalui tools rapidminer. Data diperoleh dari catatan rekam medis elektronik rumah sakit periode Oktober 2022- Oktober 2024. Lima penyakit terbanyak yang dianalisis adalah Typhoid, Diare, Abdominal Pain, Pneumonia, dan Vertigo. Data diolah melalui proses pengumpulan data, preprocessing data (pengurangan dimensi atribut, penjumlahan jumlah kunjungan perbulan lima penyakit yang dianalisi, dan transformasi data), pemodelan algoritma neural network dan operator windowing dengan parameter window size 3, dan evaluasi model menggunakan metrik Root Mean Square Error (RMSE). Hasil prediksi menunjukkan adanya pola fluktuatif untuk setiap penyakit dengan akurasi model yang tinggi, ditandai dengan nilai RMSE sebesar 7.408. kesimpulan penelitian ini menunjukkan bahwa algoritma neural network efektif dalam memproyeksikan jumlah kunjungan pasien berdasarkan diagnosa, sehingga dapat mendukung perencanaan strategi rumah sakit. Rekomedasi penelitian selanjutnya untuk mempertimbangkan faktor lain seperti faktor musiman dan demografi untuk meningkatkan akurasi prediksi.
Kata Kunci : Prediksi, Diagnosa, Data Mining.
ABSTRACT
For the sake of improving service quality, optimizing resources, and the need for a data-based health service management strategy. Given the fluctuations in the increase or decrease in patient visits that are often not accurately predicted, but to accurately know the number of future patient visits is a challenge for health institutions, especially in hospitals. This study aims to predict the number of patient visits based on disease diagnosis at dr. Bratanata Jambi hospital using a data mining approach with neural network algorithms implemented through rapidminer tools. Data was obtained from the hospital's electronic medical record records for the period October 2022-October 2024. The five most common diseases analyzed were Typhoid, Diarrhea, Abdominal Pain, Pneumonia, and Vertigo. Data was processed through data collection, data preprocessing (attribute dimension reduction, summation of the number of monthly visits of the five diseases analyzed, and data transformation), modeling of neural network algorithms and windowing operators with window size parameters of 3, and model evaluation using the Root Mean Square Error metric (RMSE). The prediction results showed a fluctuating pattern for each disease with high model accuracy, characterized by an RMSE value of 7,408. The conclusions of this study show that neural network algorithms are effective in projecting the number of patient visits based on diagnosis, so that it can support hospital strategy planning. The next recommendation of the study is to consider other factors such as seasonal factors and demographics to improve the accuracy of the predictions.
Keywords: Prediction, Diagnosis, Data Mining.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Subjects: | R Medicine > RZ Other systems of medicine |
| Divisions: | STIKES Garuda Putih > S-1 Administrasi Rumah Sakit |
| Depositing User: | SIP Fitri Suciati |
| Date Deposited: | 11 Mar 2026 02:57 |
| Last Modified: | 11 Mar 2026 02:57 |
| URI: | http://repository.stikes-garudaputih.ac.id/id/eprint/283 |

