This repository contains self-contained Recommendation System examples.

Table of Contents

• Movie Recommender-IMDB Calculation
• Movie Recommender-Parameter Filtering
• Movie Recommender-Collaborative Filtering
• Movie Recommender-Contentbased Filtering
• MovieRecommender-MatrixFactorization
• CarsRecommender-KNN
• Grocery Recommender-Apriori Algorithm
• Restaurant Recommender-Popularity Similarity Approach

---

• Movie Recommender using IMDB Calculation

  This example illustrates how to recommend movies on the basis of IMDB Formula. [Source Code]

  • Load the Dataset [movies_metadata.csv]
  • Apply IMDB Formula
  • Sort movies based on score calculated
  • Get the top n movies

---

• Movie Recommender using Parameter Filtering

  This example recommends latest movies on the basis of various filters applied on the Dataset. [Source Code]

  • Load the Dataset [movies_metadata.csv]
  • Filter by adult/Non-Adult Movie
  • Filter by year
  • Filter by Vote Count
  • Filter by Run time
  • Filter by Vote Average
  • Filter by Genres

  ---

• Movie Recommender using Collaborative Filtering

  This example demonstrates the concept of User-User Similarity and Item-Item Similarity. [Source Code]

  • Load the Dataset [u.data]

  • Create User-Movie Rating Relationship Table

  • Cater Missing Values

  • Prepare User-User Similarity Matrix Dataframe [Cosine Metric]

  • Handle diagnols

  • Create Movie-User Rating Relationship Table

  • Cater Missing Values

  • Prepare Item-Item Similarity Matrix Dataframe [Correlation Metric]

  • Load the Dataset [u.item]

  • Prepare user defiend functions to compute Recommendations.

    ---

• Movie Recommender using Contentbased Filtering

  This example demonstrates the concept of Content based Similarity (TF-IDF). [Source Code]

  • Load the Dataset [movies_metadata.csv | metadata_clean]

  • Compute TF-IDF on movie description.

  • Prepare user defiend functions to compute Content based Recommendations.

    ---

• Movie Recommender using Matrix Factorization

  This example demonstrates the concept of Matrix Factorization for recommending movies. [Source Code]

  • Load the Dataset [u.data | u.item | u.user]

  • Compute Matrix Factorization.

  • Give a sample and get the list of revevant recommendations.

    ---

• Cars Recommender using KNN

  This example demonstrates the concept of KNN for recommending movies. [Source Code]

  • Load the Dataset [mtcars.csv]

  • Compute KNN.

  • Give a sample and get the list of revevant recommendations.

    ---

• Grocery Recommender using Apriori Algorithm

  This example demonstrates the theory and implementation of Apriori Algorithm. [Source Code]

  • Install apyori library.
  • Load the Dataset [grocery.csv]
  • Picked top 2000 out of total 75000 records. Apriori is very expensive in calculation and computation.
  • Convert pandas dataframe into a list of lists.(Format for Ariori Algorithm)
  • Call apriori() function with approprite parameters.
  • Examine the results.

---

• Restaurant Recommender Using Popularity Similarity Approach

  This example demonstrates restaurant recommendations based on Rating value/Footfall.And further suggests other restaurants of similar variant. [Source Code]

  • Load the Dataset [chefmozcuisine.csv | geoplaces2.csv | rating_final.csv]
  • Compute Popularity (Rating value/Footfall)
  • Compute PearsonSimilarity.