Risk based pricing using k-means clustering.

Credit Risk Analysis is a process adapted by loan associations to evaluate the risk -based pricing for sanctioning loans. The chances of approving loans depend on the customer's credit score, which includes annual income, banking history, loan, insurance and different elements of risks such as default risk, credit correlation risk, collateral risk, credit concentration risk, rating migration risk and recovery risk. Results of the calculation allow lenders and customers to assess credit profile characteristics. Data Science allows discovering of new patterns within a complex data set which can be an efficient method to compute risk percentage by means of non – hierarchical clustering. K-means is an iterative algorithm - the idea is to use small unstructured groups of data of fixed size, assigned to clusters to be stored in the memory depending on the previous locations of the cluster centroids. It is validated that the suggested technique predicts better accuracy, is a faster and targeted approach than existing methods. [ABSTRACT FROM AUTHOR]