Download the dataset from GroupLens. Place the .zip folder in the root of this project and run database/create_database.py to create the databases.
Run your compiled Main class from the terminal (or your IDE).
Available commands:
predict <user-id> <item-id>| Predicts the rating for a user and an item.
Run weka/TestRunner to perform a k-fold cross validation.
Change how the recommender system should behave by modifying the constants in the class core/Configuration. The properties let you enable/disable features and tune parameters for the prediction.
Make sure the *.jar files in the libs folder are included as libraries in your project. Add those manually in your IDE.
Or download them by running "compile.sh".
The file DatabaseUtils.props contains the path to your database file (relative from root folder). Change this path if the database is not found when running Weka tests.
The core package contains domain independent classes and interfaces for the recommender system.
Put domain specific classes in the domain package. These classes are implementations of the interfaces in core.
The class core/RecommenderSystem is the main interface for the recommender system functionality. Main runs a command line interface that is using an instance of core/RecommenderSystsem.
Classes for evaluation are located in the weka package.