Analyzing Temporal, Spatial, and Historical Data in Rating Prediction Algorithms: A Comparative Study
Fall 2022, CSE 258: Recommender Systems and Web Mining, UC San Diego
In this project, we utilize the Google Local Reviews data set to predict the potential rating a user will give to a business. Specifically, we compare and contrast the effects of including temporal and spatial aspects of the data in our models for the task at hand.
We show that the performance of machine/deep learning models improve when we provide temporal, spatial as well as historical data. In other words, additional metadata improves the predictive power of a regression model.
Please refer to Report for more details.
Contributors
- Sanidhya Singal
- Manas Sharma
- Shyam Renjith