 |
International Society of Science and Applied Technologies |
| Zero-Inflated Binary Lasso and Ridge Regression Models with Imbalanced Data | ||||
| Author | Tzong-Ru Tsai
|
|||
| Co-Author(s) | Hsien-Ching Chen; Yuhlong Lio
|
|||
| Abstract | This study proposes three gradient descent methods to develop the estimation processes of the zero-inflated binary lasso and ridge regression models with imbalanced data. The gradient descent method with constant, momentum, and adaptive moment learning rates is used to obtain reliable estimators of the two zero-inflated binary regression models. Simulation results indicate that the proposed gradient descent methods work well. The penalized regression method can provide an efficient alternative to the expectation-maximum estimation method proposed by Chiang et al. (2024) to enhance the feature-selection performance for big data applications.
|
|||
| Keywords | Gradient descent method; lasso; ridge regression; zero-inflated | |||
| Article #: RQD2024-88 | ||||
Proceedings of 29th ISSAT International Conference on Reliability & Quality in Design |