A Multi-domain Benchmark for Personalized Search Evaluation

This dataset represents a new version to the original dataset found on Zenodo with DOI 10.5281/zenodo.6606557, tailored for Personalized Search evaluation across multiple domains. It has been generated using a fork (here) of the original experiment hosted in AmenRa's repository. The fork introduces modifications steps to the data processing steps, specifically optimized to reduce memory consumption. These adjustments enable the dataset generation process to be executed on hardware with limited resources, requiring no more than 30 GB of RAM and 16 CPU cores, and designed to complete within a timeframe of 1.5 days.