Loan Eligibility System: Exploring a Transition from Manual to Automated Processing

Loan Eligibility System: exploring a transition from manual to automated processing

Submitted as part of Business Research Methodology project

Submitted by-

Jasmine Goyal

Priyanshi Jain

Rajat Goel

Sravya Bharani M

Shitiz Gupta

Tushar Batra

Table of Contents

1. EXECUTIVE SUMMARY 2

2. INTRODUCTION 3

3. RESEARCH APPROACH 4

4. RESEARCH METHODOLOGY 6

5. RESULTS 11

6. CONCLUSION 11

7. FUTURE WORK 11

8. REFERENCES 12

1. EXECUTIVE SUMMARY

This report addresses the problem of Dream Housing Finance Company that has been losing its customers on account of delayed loan disbursements. The company would like to automate its loan eligibility process for a candidate who wants to avail a home loan. So, the two important questions that need to be answered are:

i) Does an automated loan eligibility determination model reduce the processing time of the loan disbursement process?

ii) How accurately can the automated system determine the eligibility of an applicant?

The method of analysis includes qualitative research of the existing home loan industry that includes an interview of a Home Loan Manager from the State Bank of India along with available literature on this topic. This research effectively drives home the fact that an automated system can substantially reduce the overall time required for loan disbursement.

To check how accurately an automated process can determine the eligibility of a candidate, quantitative research was taken up. A logistic regression model was built after the treatment of missing values using k-NN method. The model was built in a number of phases, dropping the most insignificant independent variable in each phase, until just significant variables remained in the model. The multi-collinearity and interaction effects between the variables were also taken into account.

The model was built on a training dataset and