Deployed a Movie Recommendation System using aws, Django, python & Machine Learning algorithms such as Collaborative Filtering Algorithms (using Matrix Factorization and Neural Networks).
The purpose was to evaluate how an existing algorithm in a movie recommender system predicts movie ratings and get an indication of how the users perceive the recommendations given by the system. The recommendations can be computed with a Low Rank Matrix Factorization (also called as SVD) algorithm that calculates both the user and movies latent features to predict ratings. Secondly, the recommendations are also computed with a revised Collaborative Filtering algorithm which makes use of Neural Networks.
- Front End Programming Tool:
- HTML, CSS, Bootstrap, jQuery, Django Template language
- Back End Programming
- Django Web Framework, Python version 3.6
- Database
- Sqlite, which is the default db used by Django
- Machine Learning Packages for Recommendation Algorithms
- Scipy, for MF (Matrix Factorization)/SVD Algorithm
- Keras for NCF (Neural Collaborative Filtering) Algorithm
- Movielens dataset
- collected by the GroupLens Research Project at the University of Minnesota.
- This data set consists of:
- 100,000 ratings (1-5) from 943 users on 1682 movies.
- Each user has rated at least 20 movies.
- Simple demographic info for the users (age, gender, occupation, zip)
-
Install Python version 3.6 from python's official site for Windows.
-
Install Django
-
Install all the Machine Learning pacages and tools, which are as follows: Tensorflow, Keras and scipy.
-
Install Django Bootstrap for front-end programming purpose.
-
To install all the packages once, use this command:
pip install -r requirements.txt
-
cd "Project Code"; Enter into the project diretory, first. load the data into the database first using the code provided in load_.py files. or just use the db file provided at this link*
-
Then, run this command to run the project:
python manage.py runserver; by default the code will run on port 8000.- If you want to change the port; run this command:
python manage.py runserver 0.0.0.0:Port_number_on_which_you want_to_run. - Also, The model.h5 file provided here should be download prior.
-
Demo has been provided at this url:
Anmol Mann, mann.anmol15@gmail.com