BINARY LOGISTIC REGRESSION ANALYSIS ON LOAN REPAYMENT STATUS
Digital technology has grown and impacted the financial industry, especially in payment methods. These days people are getting used to using cashless payment for their transaction activities. The online loan facilities provide a handy procedure and do not require a loan guarantee which makes it inte...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/65419 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Digital technology has grown and impacted the financial industry, especially in payment methods. These days people are getting used to using cashless payment for their transaction activities. The online loan facilities provide a handy procedure and do not require a loan guarantee which makes it interesting for the borrowers. Unfortunately, there are some people take would take the advantage and did not do the repayment on time. Hence, the loan issuer got a loss due to those conditions. To reduce that risk, the loan issuer is required to do some data modeling based on the characteristics of customer. Logistic regression would be the right method to predict the repayment status of the upcoming loan. The response variable is status of the loan repayment, and the predictor variables are disbursed amount, AOV, age, active period, gender, seller active, province, and product category. Multicollinearity test is used to make sure that there is no collinearity within variables. Meanwhile, the goodness of fit tests used are pseudo ????2, confusion matrix, and performance metrics. The logistic regression analysis turns out with 8 significant variables out of 19 total variables and the model has 62% accuracy, 64.1% sensitivity, and 53.5% specificity. |
|---|