Geek asks and answers

Using scikit learns to predict good website content

I have data for a year from the website. I would like to prepare a machine learning algorithm for predicting the success of new content based on certain variables (e.g. word count, publication day time, etc.).

I would like to take a new piece of data, introduce certain characteristics and get the probability of how well it will work on the site.

In addition, I would like to continue to add future data to the training set and to constantly train the algorithm so that it becomes more intelligent over time.

My question is: how should I use scikit-learn for this?

+3
source share
1 answer

, .. , .

, scikits-learn , , , .

, , :

  • SVM
  • ( scikits)
  • (Ridge, Lasso, IRLS, )
  • k

? ( 10 , 5 ). (5.1) scikits, .

. , , . , , . .

, , . , - , . , -, , k .

, scikit-learn doc : - http://scikit-learn.org/0.10/auto_examples/linear_model/logistic_l1_l2_sparsity.html#example-linear-model-logistic-l1-l2-sparsity-py - http://scikit-learn.org/0.10/modules/linear_model.html#ridge-regression

http://scikit-learn.org/0.10/user_guide.html

+3

More articles:


All Articles