Decision Support System for Indonesian Digital Bank Recommendations Using AHP-CoCoSo (Indonesian Journal of Multidisciplinary on Social and Technology)

Summary

• The study evaluates a multi-criteria decision-making system across 17 digital banks registered and supervised by OJK in Indonesia.p.1
• The hybrid decision support model was developed and tested using Ground Truth preference data from 76 user respondents.p.1
• Evaluation incorporates 3 main criteria and 14 sub-criteria spanning Application Performance, Financial Reports, and User Experience.p.1
• Individual ranking results show that Jenius at 53% and Seabank at 46% predominantly hold the primary recommendation position.p.1
• Macro performance evaluation demonstrates a macro classification accuracy of 71,36% using a Confusion Matrix.p.1
• The model achieved a Mean Absolute Error (MAE) of 4,23 rank points and a positive Spearman Rank Correlation of 0,33.p.1

Key insights

1. Jenius and Seabank cumulatively accounted for 99% of top recommendation choices across all respondent rankings.p.5
2. Seabank and Jenius achieved entry into the Top 5 recommendation zone across all 76 respondents under the AHP-CoCoSo method.p.5
3. In Ground Truth preference data, Seabank appeared 67 times and Jenius appeared 35 times within the five-largest recommendation zone.p.5
4. The Confusion Matrix evaluation reached a peak individual classification accuracy of 88,2% for 13 respondents.p.6
5. For 47 respondents, the system recommendation position shifted by less than 5 ranks from actual user choices.p.6
6. Spearman rank correlation analysis indicated positive correlation for 72 respondents, including 8 respondents with a very strong correlation between 0,6 and 1,0.p.7