Determination of Credit Decision Attributes Using Maximal Information Coefficient

Determining the right potential creditor is not an easy job for a credit analyst, who is expected to be fair and objective in making decisions and providing benefits for the bank. Analytical ability is needed to understand credit decision attributes that are important and effective in assessing potential creditors. Determining important and relevant attributes is a very important initial stage for credit analysts to analyze the submitted credit applications. This study proposes the use of the Maximal Information Coefficient method to find relevance between independent attributes with dependent attributes, while ranking independent attributes based on their relevance to the dependent attributes, as a basis for obtaining the optimal independent attributes subset. The results showed that through Monte Carlo Cross Validation, optimal attributes subset were obtained could be used (as recommendations for credit analysts) in an objective, accurate and efficient credit decision making process.