Timeline for Logistic Regression: Does my model selection process make sense?
Current License: CC BY-SA 3.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 30, 2016 at 19:54 | comment | added | RobertF | Yes, in part. Predictive modeling seeks to minimize prediction error & is frequently used for forecasting. Explanatory modeling is more concerned with explaining cause-effect paths between explanatory variables and the response rather than making future predictions. The $R^2$ statistic with linear regression is an explanatory tool for measuring goodness of fit. Predictive models minimize bias + variance combined, while with explanatory models researchers want to minimize bias in regression coefficients, even at the cost of higher variance and less predictive accuracy. | |
| Jun 30, 2016 at 19:34 | comment | added | vdiddy | Hey thanks a lot. Appreciate your help. I know about LASSO, Ridge, etc. but still a little confused about how they can enhance my response variables? Also, when you say predictive versus explanatory, are you referring to forecasting versus modeling the data as it is? | |
| Jun 30, 2016 at 19:26 | history | answered | RobertF | CC BY-SA 3.0 |