|
|
| |
| Addressing the problem of collinearity using regression methods
Download PDF
|
Sackineh Shamil Jasim, Nibras Talib Mohammed
|
Abstract:
In the field of statistics, building a model in multiple linear regression can be regarded one of the significant goals. However, when there is a problem of collinearity between explanatory variables in the research data, this situation can lead to inaccurate results in building the general linear regression model and estimating parameters. In order to overcome this problem, a method known as the "Lasso method" was used, and its results were compared with the Ridge regression method to evaluate its efficiency and accuracy in dealing with this type of problem. The results in the analysis of factors affecting thyroid function showed that the Lasso method for estimation and selection of variables also had better results in terms of the mean square error (MSE) criterion, which was less than the Ridge Regression method.
|
|
|
| | |