movie_recommendation_system

Movie recommendation system written in Python with PyTorch.

Movie Recommendation System

Source of the IMDB Movies Dataset

Usage Instructions

1. Create Validation Data

1.1 Create a validation data with the index of the row:

python3 generate_validation_data.py 33

2. Train a Model

2.1 Make sure to create the trained_models directory before training a model:

mkdir trained_models

2.2 Start TensorBoard:

tensorboard --logdir=runs

2.3 Train a model:

python3 train_model.py

3. Run Program

3.1 Run the program with the file path of the model:

python3 main.py trained_models/model_20241017_124650_93.pt

Note: It is a good idea to pick a model with the highest number of epochs.

Accuracy of the Model

A user chooses this as their favorite movie:

Movie_ID: 33
Series_Title: Joker
Released_Year: 2019
Certificate: A
Runtime: 122 min
Genre: Crime, Drama, Thriller
IMDB_Rating: 8.5
Overview: In Gotham City, mentally troubled comedian Arthur Fleck is disregarded and mistreated by society. He then embarks on a downward spiral of revolution and bloody crime. This path brings him face-to-face with his alter-ego: the Joker.
Meta_score: 59.0
Director: Todd Phillips
Star1: Joaquin Phoenix
Star2: Robert De Niro
No_of_Votes: 939252
Gross: 335451311.0

1. Result

Here is the result after running the movie recommendation program 100 times:

Accuracy of the recommendations: 81%

Note: The measure_accuracy.py script runs the program 100 times for me.

The accuracy of the model is 81%. There is certainly more work that has to be done to get it to a 90% accuracy. An accuracy above 75% is pretty good for a demo project.

sagedemage / movie_recommendation_system

readme