OSW

SIGNATURE WORK
CONFERENCE & EXHIBITION 2023

LOGISTIC REGRESSION PERFORMS BETTER THAN XGBOOST ON CREDIT SCORING PREDICTION: WHY IT DOES?

Name

Siqi Lan

Major

Data science

Class

2023

About

I'm Siqi Lan, a Chongqingnese and a DKU student from the Class of 2023, majoring in Data Science.

Signature Work Project Overview

On a dataset from the Prosper Loan company, this signature work compares the performance of Logistic Regression and XGBoost, using the same dataset from both experiments. Based on the results of this signature work, the Logistic Regression is more accurate than XGBoost, not only because of its simplicity and transparency, which makes it more suitable for applications requiring an in-depth understanding and explanation of credit decisions, such as in regulated industries like credit scoring, but also because of its high accuracy.

Signature Work Presentation Video