Example we have 100 independent variables with one classification (dependent variable, good or bad).
we run anova for each independent variable with classification variable (good or bad) and calculated F value. can we say based on F-value which independent variable is best for classification?
Not sure how ANOVA/F-test can be applied to categorical dependent variable. You have the chi square test and the pseudo-R2 in Logistic Regression (LR) which will give you some idea of model fit. The small p-value (< :0001) for the LR chi-squared statistic implies that one or more of the independent variables in the model is important for predicting the probability of dependent variable. Hope that helps.
F observed > F critical is a good proxy for significant variation between two variables
If dependent variable has 2 options (good or bad), How ANOVA/F-test is applied?. Please consider Independent sample t-test / Mann Whitney U test.
Hello Ramesh,
If your dependent variable is categorical (good or bad), how to run an ANOVA?? In this case you need a logistic regression, you can consider a multiple logistic regression as a better choice. Then you can use OR to see which of the predictors/independent variables has larger effect on the outcome/dependent variable.
Good luck!
No, Why would it? This is basically what stepwise regression does in a much more elegant algorithm. There are much better methods for variable selection e.g see the link and the papers therein.:
https://www4.stat.ncsu.edu/~boos/var.select/lasso.adaptive.html
David Booth
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA