Our recommendation system's main goal is to filter and predict only those movies that a user would like based on the individual data provided by the user. The different implementations applied to this project are the Content-Based Model and the Collaborative Filtering Model. At first, we worked on a content-based model to predict the movies for recommendation, but the downside of the model is that it predicts the top 10 movies based on the genre provided by the user and not by their preference. Based on the above reasons, which provided the same movie set for all users since it was based on genre, we moved onto the next model, which is Collaborative Filtering. This is a model used to create a recommendation based on the user's rating history (i.e., movies rated in the past). Collaborative recommender systems aggregate ratings or recommendations of objects, recognize commonalities between the users on the basis of their ratings, and generate new recommendations based on inter-user comparisons. Collaborative filtering is more efficient in this project than in the content-based model. To work on this project, take a pull or fork the code from this repository and modify it as per your requirements.
As per the data source, all the data files contain 1,000,209 anonymous ratings of approximately 3,900 movies made by 6,040 users who joined the service in 2000. There are three data set files added inside the dat_files folder for this project. The file format of the data set file is .dat format. These .dat format files are further processed and converted into CSV files and saved in the root directory, then those files are further used in the project. The reason for choosing this dataset is that it has a reasonable number of rows in each file. It also had both categorical and continuous data. It also helps us achieve our target goal of predicting movies based on user preference.
user_id :: gender :: age :: occupation :: zipcode
The data source provided the information that all demographic information is provided voluntarily by the users and is not checked for accuracy. Only users who have provided some demographic information are included in this data set.
- UserIDs range between 1 and 604 and also serves as the foreign key for User.dat and Ratings.dat file
- Gender is represented by an "M" for male and an "F" for female.
- Age is chosen from the following ranges "Under 18" to "56+" years of age:
- There are 21 different occupations collected by the data source.
movie_id :: title :: genres
- MovieIDs range between 1 and 3952 and also serve as the foreign key for the Movies.dat and Ratings.dat files.
- Titles are identical to titles provided by the IMDB (including year of release)
- Genres are pipe-separated | in the same column:
- Some MovieIDs do not correspond to a movie due to accidental duplicates entries and/or test entries.
- Movies are mostly entered by hand, so errors and inconsistencies may exist.
user_id :: movie_id :: rating :: timestamp
- UserIDs range between 1 and 604 and also serve as the foreign key for User.dat and Ratings.dat files.
- MovieIDs range between 1 and 3952 and also serve as the foreign key for the Movies.dat and Ratings.dat files.
- Ratings are made on a 5-star scale (whole-star ratings only).
- Timestamp is represented in seconds since the epoch as returned by time (2).
- Each user has at least 20 ratings.
Link - Link to the data set.
To start using this project with Git, you’ll need to install or have access to the following program.
- Tfidf Vectorizer
- Linear kernel
- Cosine similarity
- Train test split
- Pairwise distances
- Mean squared error
Git can be installed using a CLI or an executable file. The installation instructions can be found at the following link: Windows or Mac OSX
If you have not installed Python 3.9 already, the easiest method to install both the programs is by installing anaconda The following link provides a graphical installer link for both Windows and Mac OS Link
If you have already installed Python 3.9 and are an advanced user, you can install Jypter Notebook on terminal by following the steps below.
pip3 install jupyter
In Terminal, navigate to the respective directory where you want to clone this repository and run the following command.
git clone https://github.com/Praveen271195/Movie-Recommedation-System-DRAFT
If you have forked this repository, use the following code but replace the [username] in the link with your git username.
git clone https://github.com/[username]/Movie-Recommedation-System-DRAFT
If you have installed the Jypter notebook via anaconda, you can run the notebook directly by double clicking on the Jypter notebook icon on the start menu on Windows or Mac, going to the app drawer and selecting the Jypter notebook.
If you have installed Jypter notebook via Python, you can go to Terminal and type the following code:
jupyter notebook
Both the processes will open up the Jypter notebook environment in a default web browser.
In the application window, there will be a file explorer. Navigate to the respective folder where you have cloned the project and select the " Movie Recommendation system - DSCI Final Project.ipynb " to load up the notebook.
If you want to run this project on Google Colab, Navigate to the following link
In the application window, click on File->Open Notebook from the menu tab or press Ctrl + O. On the open popup model, click on upload file and go to the respective folder where this project is cloned and select the " Movie Recommendation system - DSCI Final Project.ipynb " so that it can be uploaded to Google Colab.
- In Menu -> select Cell -> Run all
For content-based model search for the following code snippet like mentioned below.
content_based_test_list = ["Jumanji (1995)", "Pinocchio (1940)", "Walking Dead, The (1995)", "Othello (1995)"]
For collaborative-filtering model search for the following code snippet like mentioned below.
user = [
{'title':'Breakfast Club, The (1985)', 'rating':4},
{'title':'Toy Story (1995)', 'rating':2.5},
{'title':'Jumanji (1995)', 'rating':3},
{'title':"Pulp Fiction (1994)", 'rating':4.5},
{'title':'Akira (1988)', 'rating':5}
]
Neither the University of Minnesota nor any of the researchers involved can guarantee the correctness of the data, its suitability for any particular purpose, or the validity of results based on the use of the data set. The data set may be used for any research purposes under the following conditions:
-
The user may not state or imply any endorsement from the University of Minnesota or the GroupLens Research Group.
-
The user must acknowledge the use of the data set in publications resulting from the use of the data set (see below for citation information).
-
The user may not redistribute the data without separate permission.
-
The user may not use this information for any commercial or revenue-bearing purposes without first obtaining permission from a faculty member of the GroupLens Research Project at the University of Minnesota.
If you have any further questions or comments, please contact GroupLens grouplens-info@cs.umn.edu.
To acknowledge use of the dataset in publications, please cite the following paper:
F. Maxwell Harper and Joseph A. Konstan. 2015. The MovieLens Datasets: History and Context. ACM Transactions on Interactive Intelligent Systems (TiiS) 5, 4, Article 19 (December 2015), 19 pages. DOI=http://dx.doi.org/10.1145/2827872