This is a final project of EN.601.429/629 Functional Programming in Software Engineering in Fall 2022.

An Interactive command-line application built in OCaml for movie recommendation. Use three modes based on the three types of recommendation systems to generate the recommendation list:

• Demographic Mode

  It offers generalized recommendations to every user based on movie popularity, which can be treated as a top list among all movies in our movie library. Users don’t need to input their movie preference list, and every user will get the same ordered list.

• Content Based Mode

  It provides similar movies. First genreate TF-IDF vector for feature overview, and counter vector for features keywords, genres, casts, and director, and then combine these two vectors to calculate cosine similarity for movie pairs. Users need to input their movie preference list, and will get a similar recommended movie list.

• Collaborative Mode

  It provides recommendations based on the rating data. Use biasSVD to predict rating for movies. Users need to input a rating list which contains the movie and corresponding rates.

1. Install python library sklearn (See the official instructions)

   $pip3 install -U scikit-learn

2. Ensure that csv , yojson, owl, sklearn , inquire , ppx_deriving, ounit2, bisect_ppx and its dependencies are installed by running:

   $opam install csv yojson owl sklearn inquire ppx_deriving ounit2 bisect_ppx

3. Build project in the project directory, it will generate an executable file main.exe

   $dune build

   Then run the program with the following command

   $./_build/default/src/main.exe

   Or directly use the following command to build and run the program

   $dune exec ./src/main.exe

• movie.ml - The movie data related data type Movie.t and functions
• rating.ml - The rating data related data type Rating.t
• data_preprocessing.ml - Parse csv file and load data, call load_movie_data (credits_filename) (movie_filename) to load movie dataset and load_rating_data (rating_filename) to load rating dataset
• utils - Common used functions
• demographic_mode.ml - Main logic of demographic mode, call get_recommendations (movie_list:Movie.t) (n:int) to get recommended movie list
• content_based_mode.ml - Main logic of demographic mode, call get_recommendations ~(title:string) ~(n:int) ~(movie_list:Movie.t) to get recommended movie list
• collaborative_mode.ml - Main logic of collaborative mode, call get_recommendations ~(movie_list:Movie.t) ~(rating_list:Rating.t) ~(ur_list:Rating.t) ~(n:int) to get recommended movie list
• main.ml - Main logic for interactive commend line interface

• Getting Started with a Movie Recommendation System | Kaggle