Implement linear regression about golearn HOT 6 CLOSED

@sjwhitworth as discussed, MIT licence added to our multivariable regression repo. Feel free to add what you like, or i can help out as you get the interfaces into shape...

https://github.com/sajari/regression

Also, this contains an example parsing csv's for use. It assumes a header row and float64 for data, but should be quick to smash into something more generic.

from golearn.

Interfaces are designed now - can you just implement the methods defined here?
@marcoseravalli - we've got a start on regression now - can you help integrate?

https://docs.google.com/a/hailocab.com/document/d/1x21Y-g1rga0LTwC_LnKHi0y7RjFzd2Il7YB47rp7kTA/edit

For linear regression, it'd be Fit(), Predict() and Score(). I don't think you have to change too much. I can help integrate as well.

from golearn.

sorry but for the time being I cannot find the time to work on this, it would be probably better to assign it to someone else

from golearn.

What's the status of this? Has it been implemented in the meanwhile?

from golearn.

This is implemented in the linear package

from golearn.

In fact, I have implemented liblinear already, the Golang wrapper of LIBLINEAR (C library). With it, the linear algorithms (logistic regression, linear svm etc.) are ready for use. We can further wrap the cross validation if needed.
Please feel free to initiate the integration. 😄

from golearn.