RAW_interactions = links RAW_recipes = links

• We are in the process of developing a personalized food product recommendation software. This software will suggest food recipes to users based on their interactions, preferences, similarities, and the popularity of products. We intend to use a hybrid approach, combining various recommendation techniques. This system aims to provide users with relevant recipe suggestions tailored to their preferences and behaviors.

• We have implemented three types of filtering in our system:

  1. Content-Based Filtering: This will filter the data according to particular user’s interaction and preferences by using tags and description present in data frame.
  2. Collaborative-Based Filtering: This will filter out the data got from content-based filtering according to the similar users who reviewed that particular dish.
  3. Popularity Based Filtering: This will filter the most popular recipes got from content based filtering according to the ratings.

• We have now assigned weights to each set of filtered data, with greater weight assigned to more preferred dishes. Using a weighted hybrid approach, we will determine the recommended dishes. In this approach, different weights are assigned to the filtered data from the three models. This method aims to provide a balanced and personalized recommendation by considering the preferences and interactions of users with varying degrees of importance.

• We are assigning the highest weight to content-based filtering, considering it as the most crucial factor. Next in priority is collaborative-based filtering, followed by popularity-based filtering. This weighted sequence ensures that the recommendation system places utmost importance on content-based insights while also incorporating collaborative and popularity aspects. As a result, we anticipate achieving enhanced accuracy in our recommendations..

• This will give us most recommended food recipes.