1 if the data set must contain strong characteristics, of course, there are other weak characteristics, independent character, so if you don't do feature selection, will have an impact on the prediction accuracy,

I ask this question, because I think there will be no influence, that since there won't be affected, regardless of dimension disaster, why to do feature selection,

3, if in the data set was not strong features, so no matter how to choose, also does not solve the problem,

I don't know I'm right, I consult, thank you,

CodePudding user response:

1. The effect should be not big, complex model will automatically find strong features, such as neural networks
2. Sometimes colinear problems
3. To manually build useful feature, feature of engineering