/recommender-systems

Primary LanguagePython

Recommender Systems

How to use?

  1. Install Docker from: https://www.docker.com/

  2. Start the Docker daemon and navigate to the root folder

  3. From terminal run make server

  4. Now you will have two endpoints open at the url http://0.0.0.0:5000/

    • /top-items - returns the most voted items

    • /collab-filt - returns the most suggested products for the user

How to use the end points?

You have to make a POST request to the end points, the body containing JSON data.

/top-items end point

For the /top-items end point, you have to send a JSON which has a key items and the value for the key is an array which contains also JSON objects. Each object is an item from the dataset, and has the following values in it.

    - id - item id integer
    - name - item name string
    - vote_average - vote average for the item, float
    - vote_count - number of times the item was voted, integer
    - description - description of the item, string

It returns the same objects in a JSON array, but ordered according to the criteria. Then you only have to display them.

/collab-filt end point

For the /collab-filt end point, you have to send a JSON which has two keys.

    1. *ratings*, which has as value an array which in fact contains objects of the JSON form

            - userId - integer
            - itemId - integer
            - rating - float between 1 and 5
    
    2. *recomm_user_id*, is the ID of the current user to which we want to make a recommendation

It returns a JSON object which has a key result and value an array of item IDs. Those item ID's are the recommendations which you have to show.