Modification to Standard Logistic Regression Model to Improve the Performance of Drug Discovery

The recent years have witnessed an increase in attention to improving the prediction results of drug side-effects and therapeutic indications. Most of such studies have used standard regression model for prediction which suffers from overfitting. In this paper, a solution is offered for overfitting problem. This solution is based on regularization which has improved the prediction results of standard regression model. Experimental results show that the parameters of sensitivity, specificity, and AUROC in suggested method improve compared with standard regression model.

کلیدواژه ها:

therapeutic indication ، side-effect ، sensitivity ، specificity ، AUROC ، regularized logistic regression.