docs: add model selection notes

docs/model selection.md

@@ -0,0 +1,6 @@
+
+* Random forest are good when there are a lot of categorical features and many of them are not relevant.
+* Nonparametric methods are good when you have a lot of data and no prior knowledge and when you do not want to worry to much about choosing the right features (As it is expensive to run make sure there is less then 20 features or so).
+* Logistic regression when the data is linearly separable or can be made into it by feature engineering.
+* Support Vector machines are good when 
+* When having noisy and much data use neural networks