Loan Approval System through Customer Segmentation Using Big Data Analytics and Machine Learning

Abstract

Data Science and big data analytics, is a field that assesses, and extricates data from excessively huge informational collections that are complex to be otherwise managed by conventional information handling applications and programming. Among all businesses, the financial area has probably the biggest consumption of big data and information science strategies. It is relied upon to change the large credit and personal loan industry in the banking sector, in view of the bits of knowledge and insights it generates about borrowers who are hard to endorse. The aim for this project study is to evaluate the credit's default likelihood and risks through studying the authentic dataset obtained from the financial organizations with various customer attribute variables and then ordering the people into one of two classes: (a) Uncertain/Risky (b) Eligible/ seeming to satisfy the repayment in full. The strategy towards making this characterization requires understanding the consumers and their past record as a consumer through data exploration and correlations between the important features and loan status followed by applying various machine learning algorithms like random forest, logistic regression and support vector mean to estimate the accuracy levels of the trained dataset model to process future loans. The project can quicken the credit creation cycle, mitigate risks, and render loan approval quickly. Along these lines, analytics can quickly survey the dangers of giving a credit and help settle on the choice of sanctioning a specific request for loan in a capable way.

Keywords:  big data analytics, machine learning, banking sector, credit,  feature importance, customer characterization.