Consider a platform for content recommendation based on the user history. The contents are books and articles and by history I mean what the user has read, what he has shared and so on. I know that there are many studies on how to train a recommendation system, however I don't know what to do to "start" (the platform has not been launched yet and the dataset is empty). Are there techniques that consider an initialization phase where no data is available?


This is often called the cold start problem.

There are many options to initialize:

  • Random suggestions
  • Domain expert suggestions
  • Most popular suggestions from another platform
  • Content-based recommendation

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.