Abstract: This paper is used to cluster the various components of a bank customer details and segregate potential customers eligible for loan. The application of this technique helps the banker to scale the potentiality of their customers and take necessary steps to decide for loan approval. The classic difficulty of recognizing the customerís potentiality is solved using this thesis and helps them to judge the good will of their customers.Identifying the eligible customers for loan in this modern society is complex. The identified customer must be able to repay his loan in the proper installments throughout the tenure. The Banker can identify the loyalty of the customer with the help of this thesis. After completing the collection of data, it is clustered according to the monthly salary, movable and immovable assets. Since K-Medoid algorithm is an unsupervised algorithm, we specify the number of clusters.

Keywords: K-Medoid, Clusters, Unsupervised Algorithm, Potential Customers, Loan Installment, Goodwill.