Recommendation-System

A system that can be used to build user profiles and accordingly recommend items to buy/watch/read that matches their users’ interests

• Recommend an item to the user based on their previous purchases if the current cart is empty
  
• If there are items already added to the users cart, recommend an item to the user based on their previous purchases and the current items in the cart
  
• Recommend an item to the user based on the intersection between the items they previously purchased and other users’ purchases
  

Running the tests

Given the below toy dataset where users is a list of the users, items is a list of the items that the users can buy and purchases is a list mapping each user to a list of all his/her previous shopping carts.
users = ["A", "B", "C", "D", "E"]

items = ["suit", "dress", "shoes", "T−shirt", "Jacket", "skirt", "shorts", "shirt", "trousers", "scarf", "mp3player", "TV", "LCDscreen", "headphone", "laptop", "keyboard", "mouse", "cellphone", "headphone", "earphone", "milk", "cheese", "bread", "chocolate", "meat", "flour", "sugar", "oil", "tomatoes", "chicken", "yogurt", "cereal", "beans", "fool", "eggs"]

purchases = [("A", [["dress", "shoes"], ["milk", "cheese", "eggs"]]), ("B", [["earphone", "mouse", "laptop"], ["mp3player"]]), ("C", [["bread", "milk"], ["shoes"]]), ("D", [["milk", "meat", "chicken", "yogurt"], ["beans", "cereal", "flour", "sugar"], ["tomatoes", "oil", "chicken"]]), ("E", []) ]

Functions:

recommend

recommend :: String -> [String] -> String
The function recommend takes a user and his/her cart and recommend to the user an item accordingly

> recommend "A" ["earphone"]
"laptop"
> recommend "E" ["dress"]
"cereal"
> recommend "A" ["earphone"]
"tomatoes"

recommendEmptyCart

recommendEmptyCart :: String -> String
When the users cart is empty, the function recommendEmptyCart could be used to recommend an item to the user by 50% based on previously purchased items and 50% based on the intersection between the items purchased by the current user and other users.

> recommendEmptyCart "E"
"laptop"
> recommendEmptyCart "E"
"chicken"
> recommendEmptyCart "A"
"tomatoes"

recommendBasedOnItemsInCart

recommendBasedOnItemsInCart :: String -> [String] -> String
The function recommendBasedOnItems is used to recommend an item to the user based on the items currently in the user’s cart and the previously purchased items by 50% and the other 50% based on the intersection between the items purchased by the current user and other users.

> recommendBasedOnItemsInCart "A" ["dress"]
"shoes"
> recommendBasedOnItemsInCart "A" []
"sugar

recommendBasedOnUsers

recommendBasedOnUsers :: String -> String
The function recommendBasedOnUsers is used to recommend an item to the user based on the intersection between the items purchased by the current user and other users.

> recommendBasedOnUsers "B"
"meat"
> recommendBasedOnUsers "A"
"meat"
> recommendBasedOnUsers "A"
"chicken"
> recommendBasedOnUsers "C"
"earphone"

Acknowledgment

This project outline was designed by the MET department at the German University in Cairo link.