辅导machine learning、讲解c/c++，Java、辅导Python程序
- 首页 >> 其他 Tips:
> If the distribution of predictors is very skewed or others rather than normal distribution, we
should do some reasonable transformation to make its distribution closer to normal distributio
n
> the transformation should be done before the normalization
> Dummy variables and label encoders do not need transformation
> Left skewed should be transferred into the right skewed by making values negative. And yo
u can apply log transformation to it.
> square root can be considered
> more useful tips could be found https://machinelearningmastery.com/how-to-transformdata-to-fit-the-normal-distribution/
or other useful methods you can applied with supporting
> If the distribution of predictors is very skewed or others rather than normal distribution, we
should do some reasonable transformation to make its distribution closer to normal distributio
n
> the transformation should be done before the normalization
> Dummy variables and label encoders do not need transformation
> Left skewed should be transferred into the right skewed by making values negative. And yo
u can apply log transformation to it.
> square root can be considered
> more useful tips could be found https://machinelearningmastery.com/how-to-transformdata-to-fit-the-normal-distribution/
or other useful methods you can applied with supporting