Tree-based methods are the only ones that are invariant towards feature
scaling, do DecisionTree*, RandomForest*, ExtraTrees*, Bagging* (with
trees), GradientBoosting* (with trees).
For all other algorithms, the outcome will be different whether you
scale your data or not.
For algorithms like nearest neighbors, I would not say they require
scaling, but scaling will change the result.
It is then a question on whether you think the range of your features is
meaningful or arbitrary.
I don't think there is currently a chart on the complexity of
algorithms, thought it would be cool to add.
On 06/04/2015 07:14 AM, Yury Zhauniarovich wrote:
Hello everyone,
I have the following question. In general, as a rule of thumb features
need to be scaled using min-max scaling or z-score standardization
before being used in ML algorithms. However, it is not always possible
to perform this procedure (e.g., in cases when you do not have all the
data, or you do not have enough resources to perform this operation).
At the same time, some classification algorithms do not require data
scaling to operate correctly (e.g., Random Forest classifier). Correct
me if I am wrong in this assumption.
If it is possible, could you name please classification algorithms
that do not require feature scaling and those which require?
And one more question. Have you ever seen the comparison of algorithms
in terms of their speed and memory consumption (maybe, there is such
comparison for algorithms in scikit)? Where I can find the information
which algorithms are more greedy and which are not?
Sorry if my questions seem to you too basic but I am in the beginning
of my way. Thank you in advance!
------------------------------------------------------------------------------
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
------------------------------------------------------------------------------
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
