I want to know what is the best way to find the most suitable k in knn algorithm for Item-based collaborative filtering . I want to make a recommendation system for an application to recommend the most similar items (products) for the product which user liked the most based on rating values . The application is an ecommerce application and the products are published by the users (as social apps) the data which used as input to the algorithm is not huge. It depends on how many products the users post in the application.

note: if it makes sense, I used cosine similarity to find the similarity between items.

  • $\begingroup$ This article has a nice section, titled "How do we choose the factor K?", which will help you understand how K affects your clustering, and how to choose the right K for your specific task. analyticsvidhya.com/blog/2018/03/… $\endgroup$ Dec 11 '20 at 7:48
  • $\begingroup$ @user3667125 thank you ^_^ $\endgroup$
    – roro roor
    Dec 11 '20 at 9:17

Your Answer

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

Browse other questions tagged or ask your own question.