The objective of this research is to assess financial institutions across various countries and regions from 2011 to 2020 using hybrid machine learning methodologies. Machine learning is employed for data analysis and prediction, proving particularly effective for extensive datasets. This paper analyzes a real bank credit dataset to examine the functionality of bank credit. The random forest method identifies Net Interest Income (gross profit and loss) as the most influential factor. Using Fuzzy C-means, we categorize the data into five clusters across the studied years. With Cluster-Induced Ordered Weighted Averaging (CIOWA), 2013 is identified as the best-performing year. This study contributes to the field by applying hybrid machine learning methods to forecast the future performance of financial institutions. Additionally, a comprehensive literature review on related issues is incorporated into this model.