Early Detection of Failed Bank Through Analysis of Financial Ratios and Bank Shareholders Ratios Using Data Mining For Rural Banks

Abstract

This study aims to determine the effect of 8 bank financial ratios such as BOPO (operational efficiency ratio), CAR (Capital Adequacy Ratio), NPL (Non Performing Loan), ROA (Return On Assets), CR (Cash Ratio), KAP (quality of productive assets), PPAP (provision for loan losses) and LDR (Loan Deposit Ratio) and another ratio, namely Bank’s Shareholder ratio towards bank predictions whether a rural bank will be declared as failed bank or not. Eight financial ratios and another ratio that comparing BOD and BOC to Bank's Shareholders can be obtained from quarterly rural bank’s financial reports that have been published on the IFSA website from 2014 until 2018. The data in this research is approximately 1000 rural banks for training dataset. The method to predict rural bank become failed bank is data mining. The training dataset used is an imbalanced dataset. In order to be balanced, the SMOTE method is used. The balance dataset was then analyzed with the data mining process. The data mining methods used are KNN and Naïve Bayes, both are classification method.