Comparing Linear Regression to Shrinkage Regression Algorithms (RR, Lasso, El Net) Using PTSD Patients’ Data

The purpose of this research was to introduce the alternative model of regression algorithms and having it compared to linear regression. To do this, we need to use modern algorithms such as Ridge, Lasso, and Elastic net regression in which precision is maximized by regularizing the cost function. In this paper theoretical basis and practical implications have been explained.. The target population was patients diagnosed with Post Traumatic Stress Disorder (PTSD) in 2020 for the comparison. 97 PTSD patients (73 females and 24 males) in Tehran were measured in 8 variables related with the intensity of the trauma re-experience. The linear regression, Ridge, Lasso, and Elastic Regressions were used with R software. The results indicated that compared to linear regression, Elastic. LASSO and Ridge explained more variances and had more R square and less MSE respectively. When the main assumptions of Linear regression are not met, using shrinkage regressions seems to be reasonable and accurate.