PEMBANGUNAN MODEL PENDETEKSI RISIKO PREEKLAMSIA PADA PASIEN RS ISLAM JAKARTA PONDOK KOPI DENGAN MENGGUNAKAN TEKNIK DATA MINING

• Main Author: Ilham, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/68053
• Institution: Institut Teknologi Bandung

One indication that can be used to measure the performance of the hospital is to improve health services in order to prevent death in patients. One of the causes of death in patients is abnormalities in pregnant women caused by preeclampsia. Rumah Sakit Islam Jakarta Pondok Kopi (RSIJPK) is one of several private hospitals located in East Jakarta, DKI Jakarta. Therefore, RSIJPK needs to improve the quality of services offered to patients in order to compete in a competitive environment by implementing a system that can detect preeclampsia using data mining techniques. The study was conducted using six data mining classification algorithms using 109 obstetric clinic patient data at RSIJPK consisting of 48 preeclampsia positive patients and 61 preeclampsia negative patients. The input features used as attributes for the detection of preeclampsia were obtained based on the results of a literature study and the results of consultations with obstetricians. The diagnosis of preeclampsia risk is divided into two different classes, namely positive diagnosis and negative diagnosis. Based on the results of the model evaluation, logistic regression was obtained as the algorithm that has the best performance in detecting preeclampsia in RSIJPK pregnancy patients with an accuracy value of 98% with a precision level of 100%. The input features are reduced to 12 predictor features with 1 target feature. This study proposes RSIJPK to implement a preeclampsia detection system that uses a logistic regression algorithm so that it can be utilized by obstetricians, medical personnel, and patients so as to increase the hospital's competitive advantage.