Recommendation Systems

• Association Rule Based Recommender System (Armut)
• Hybrid Recommender System (MovieLens)
  • User-Based Recommendation
  • Item-Based Recommendation

Association Rule Based Recommender System using Armut's Dataset

GitHub Repository:

Business Problem
Armut, Türkiye's largest online service platform, brings together service providers and those who want to receive service.
It provides easy access to services such as cleaning, modification and transportation with a few touches on your computer or smart phone.

It is desired to create a product recommendation system with Association Rule Learning by using the dataset containing the service users and the services and categories these users have received.

Dataset Story
The dataset consists of the services customers receive and the categories of these services. It contains the date and time information of each service received.

Variables

• UserId: Customer ID
• ServiceId: Anonymized services belonging to each category
  (Example: Sofa washing service under the cleaning category)
  A ServiceId can be found under different categories and refers to different services under different categories
  (Example: The service with CategoryId 7 and ServiceId 4 is honeycomb cleaning, while the service with CategoryId 2 and ServiceId 4 is furniture assembly)
• CategoryId: Anonymized categories.
  (Example: Cleaning, transportation, renovation category)
• CreateDate: The date the service was purchased

Requirements

• mlxtend==0.22.0
• pandas==1.5.1

Hybrid Recommender System using MovieLens Dataset

• User-Based Recommendation
• Item-Based Recommendation

GitHub Repository:

Business Problem
Make a guess for the user whose ID is given, using the item-based and user-based recommender methods.
Consider 5 suggestions from the user-based model and 5 suggestions from the item-based model and finally make 10 suggestions from 2 models.

Dataset Story
The dataset was provided by MovieLens, a movie recommendation service. It contains the rating scores for these movies along with the movies. It contains 20.000.263 ratings across 27.278 movies. This dataset was created on October 17, 2016. Includes 138.493 users and data from 09 January 1995 to 31 March 2015. Users are randomly selected. It is known that all selected users voted for at least 20 movies.

Variables

• movieId: Unique movie ID
• title: Movie title
• genres: Movie genre

• userId: Unique User ID
• movieId: Unique Movie ID
• rating: Rating given to the movie by the user
• timestamp: Review data

Requirements

• pandas==1.5.1