Which type of regression focuses on selecting predictors based on statistical significance and predictive power?

Stepwise multiple regression is a statistical method that is specifically designed to select predictors based on their statistical significance and predictive power. This approach involves an iterative process where the model is built by adding or removing predictors based on certain criteria, such as the p-value.

In stepwise multiple regression, the process begins with no predictors in the model or a set of candidate predictors, and then predictors are included or excluded based on their contribution to the model's explanatory power. This often involves assessing the fit of the model and determining if the inclusion of additional variables significantly improves the model's ability to predict the outcome variable. This method is particularly useful when dealing with a large number of potential predictors, allowing for a more streamlined model that focuses on the most relevant variables.

Other types of regression, such as logistic regression, non-linear regression, and simple linear regression, either focus specifically on outcomes with particular characteristics (like binary outcomes in logistic regression) or do not utilize this iterative selection process to determine which predictors to include based on their significance and efficacy. Thus, stepwise multiple regression stands out for its systematic approach to predictor selection.