The recommendations should be based on the products consumer has searched on other sites like Google.

This basically means, that recommendations have to be made to the user based on his/her search history. No other information is available for a user.

Has this sort of thing already been implemented? (preferably in Python)

The approach I have thought of is removing stop words and extracting keywords from a search query, based on some criteria. I am falling short at the implementation.

  • 1
    $\begingroup$ Have a look on GitHub Projects! $\endgroup$ – Aditya Dec 28 '18 at 4:27

To answer your 1st question,I am not sure if a recommendation system based on search history has been implemented or not.This approach does sound cool.

Secondly, yes there are various algorithms to extract phrases from text. Phrase extraction and Text Summarization are one of the most important aspects in development of chatbots for conversational AI. To give you an example, there's a good library for this purpose, called PyTextRank. Go through this link for better understanding of how this works.

Once you have a ranked summary of phrases for the given user's search history, you will be able to generate the list of item/s , that person is interested in. This is a good start for building the person's unique feature vector.
After this process, you can incorporate normal recommendation algos like Collaborative Filtering , etc.

Hope this helps :)


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.