Basics of Recommender Systems: Study of Location-Based Social Networks
This is a simple tutorial to explore the basic of recommender systems. Here, we implemented three basic methods in recommender system, User-based Collaborative Filtering, Item-based Collaborative Filtering, and Sigular Value Decomposition (SVD).
We have two Jupyter notebooks, Data Preprocessing and Recommender Systems Algorithms. In the first one, we read the dataset and preprocess it. Then, in Recommender Systems Algorithms, we implement the basic methods and compare them.
The dataset collect from the Foursqaure by the following paper. The dataset folder includes the original dataset and preprocessed_data includes the dataset after pre-processing. The dataset has 2321 users, 5596 locations (POIs), and 151589 check-ins.
Yuan et al., Time-aware point-ofinterest recommendation, SIGIR, 2013.
We put the PDF version of the notebooks in the PDFs folder.
Please feel free to contact by rahmanidashti@gmail.com, if you require any further information.