Features enhancement of linear regression methods: “statsmodels” and “scikit-learn”

[hhan14]

Features enhancement of linear regression methods: “statsmodels” and “scikit-learn”

• Concept: Compare linear regression methods, “statsmodels” and “scikit-learn”, and abstract/combine advanced features of those two.
• Objective: To obtain maximized features out of two different approaches on linear regression.
• Why valuable: Those two methods have advanced features in terms of functions over the other in comparison. By adding those features, we don`t need to leave anything to be desired from each method.
• Features of software function: (1) abstract and add automatic ‘print coefficiency’, and ‘summary’ function to “scikit-learn”, and (2) add constant of intercept as default to “statsmodels”

[EHWUSF]

I'd say that adding things to statsmodels is slightly out of scope for this project, but augmenting the scikit-learn linear regression features certainly fits.

Please be more specific about just what specific features you propose to support, though-- it sounds like you want to augment scikit-learn linear regression with some 'print coefficiency' thing from statsmodels, but on the face of it, it sounds to me like something scikit-learn already does, and if I try google searching for "statsmodels print coefficiency" I don't really get any more clarity. And I'm not familiar with the statsmodels summary feature, so that should be explained here also.