Back to Search
Start Over
Efficient bankruptcy (credit cards) prediction using logistic regression over linear regression with improved accuracy.
- Source :
-
AIP Conference Proceedings . 2024, Vol. 3161 Issue 1, p1-7. 7p. - Publication Year :
- 2024
-
Abstract
- To enhance the precision of bankruptcy prediction by implementing Logistic Regression, as opposed to Linear Regression, with a particular focus on efficiency. The study involves a classification approach where Logistic Regression is applied to a sample of 20 cases, while a separate sample of 20 cases is used for Linear Regression. The study configured a Logistic Regression model with specific parameters (alpha, g-power, beta) and aimed to assess its performance with a 95% confidence interval (likely to ensure the results are statistically significant). They compared this to a Linear Regression model. The Logistic Regression achieved a higher accuracy (95.1%) compared to Linear Regression (88.8%) for this task. The mean accuracy detection stands at 0.660 (±2SD), with a significance value of 0.00 (p<0.05). Thus, it is evident that a statistically significant distinction exists between the two groups. The assertion is correct, and it has been conducted through an Independent Sample T-test. The investigation into bankruptcy, the Logistic Regression had greater accuracy (95.3%) while in comparison with the Linear Regression (LR). This research proposes a novel Logistic Regression approach that addresses the limitations of previous studies and achieves higher accuracy. This method empowers users to make more accurate predictions. [ABSTRACT FROM AUTHOR]
- Subjects :
- *CREDIT cards
*REGRESSION analysis
*CONFIDENCE intervals
*SELF-efficacy
*BANKRUPTCY
Subjects
Details
- Language :
- English
- ISSN :
- 0094243X
- Volume :
- 3161
- Issue :
- 1
- Database :
- Academic Search Index
- Journal :
- AIP Conference Proceedings
- Publication Type :
- Conference
- Accession number :
- 179375149
- Full Text :
- https://doi.org/10.1063/5.0229248