Fitriawati, Eka Nur (2026) Prediksi Z-Score sebagai Indikator Kestabilan Perbankan Syariah di Indonesia Menggunakan Support Vector Regression. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5002221121-Undergraduate_Thesis.pdf.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5002221121-Undergraduate_Thesis.pdf.pdf - Accepted Version Restricted to Repository staff only Download (12MB) | Request a copy |
Abstract
Kestabilan perbankan syariah merupakan aspek penting dalam menjaga stabilitas sistem keuangan nasional. Penelitian ini bertujuan untuk mengembangkan model prediksi kestabilan Bank Umum Syariah dan Bank Perekonomian Rakyat Syariah di Indonesia menggunakan metode Support Vector Regression berbasis indikator Z-Score. Data penelitian mencakup indikator keuangan internal perbankan syariah serta variabel makroekonomi dalam bentuk data time series bulanan periode 2010–2024. Tahapan penelitian meliputi pra-pemrosesan data, pembentukan fitur melalui feature engineering, pengujian empat jenis kernel SVR yaitu linear, polynomial, RBF, dan sigmoid, serta optimasi parameter model menggunakan Grid Search. Evaluasi performa dilakukan menggunakan Walk-Forward Validation dengan metrik RMSE, MAE, dan MAPE. Berdasarkan hasil evaluasi, model terbaik diperoleh menggunakan kernel polynomial pada BUS dan kernel RBF pada BPRS. Model terpilih kemudian digunakan untuk melakukan prediksi kestabilan 12 bulan ke depan menggunakan strategi Direct Forecasting. Hasil penelitian menunjukkan bahwa model SVR mampu memberikan performa prediksi yang baik. Berdasarkan analisis SHAP, variabel yang paling berkontribusi terhadap hasil prediksi meliputi faktor ekonomi berupa BI 7-Day Reverse Repo Rate, faktor internal perbankan berupa Return on Asset pada BUS dan Multi Financing Facility pada BPRS, serta fitur historis Z-Score dan rolling mean.
========================================================================================================================================
Islamic banking stability is an important aspect in maintaining the stability of the national financial system. This study aims to develop a stability prediction model for Islamic Commercial Banks (BUS) and Islamic Rural Banks (BPRS) in Indonesia using the Support Vector Regression (SVR) method based on the Z-Score indicator. The dataset includes internal financial indicators of Islamic banking and macroeconomic variables in the form of monthly time series data for the period 2010–2024. The research stages consist of data preprocessing, feature engineering, testing of four SVR kernel types, namely linear, polynomial, RBF, and sigmoid, as well as model parameter optimization using Grid Search. Model performance evaluation was conducted using Walk-Forward Validation with RMSE, MAE, and MAPE metrics. Based on the evaluation results, the best model was obtained using the polynomial kernel for BUS and the RBF kernel for BPRS. The selected models were then used to forecast banking stability for the next 12 months using the Direct Forecasting strategy. The results indicate that the SVR model provides good predictive performance. Based on SHAP analysis, the variables that contribute most to the prediction results include economic factors represented by the BI 7-Day Reverse Repo Rate, internal banking factors represented by Return on Assets (ROA) for BUS and Multi Financing Facility for BPRS, as well as historical features such as Z-Score and rolling mean.
Available Versions of this Item
- Prediksi Z-Score sebagai Indikator Kestabilan Perbankan Syariah di Indonesia Menggunakan Support Vector Regression. (deposited 29 Jul 2026 02:11) [Currently Displayed]
Actions (login required)
 |
View Item |