#TITLE_ALTERNATIVE#
In analyzing the problem of regression, important things done before the problem is modeled by considering a subset regression. Generally have pretty much of a problem so that the predictor variables in the model it will be difficult. However, if the variables derived little, could cause the results...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/18149 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | In analyzing the problem of regression, important things done before the problem is modeled by considering a subset regression. Generally have pretty much of a problem so that the predictor variables in the model it will be difficult. However, if the variables derived little, could cause the results of modeling regression inaccurate. Thus the need for a way to select a subset of variables predictor. Thus the need for a way to select a subset regression. The selection of appropriate subset will produce a good model that can present problems in the model.
Quite a lot of the methods used in selecting a good subset regression, including; method and the backward and stepwise method. Backward method selects a subset of regression with respect to the partial F value of each predictor variable, while the stepwise method than taking the value of the partial F also consider partial correlation.
After selecting subsets regression sometimes need to pay attention to the problem of multicollinearity in the predictor variable, it aims to maintain the assumption that each predictor variable regression should not be related to each other. Ridge regression a good way to solve the problem multicollinearity. |
|---|