Revisiting Alternative Experimental Settings for Evaluating Top-N Item Recommendation Algorithms
This work addresses evaluation challenges in recommendation systems for researchers and practitioners, but it is incremental as it revisits existing settings rather than introducing new methods.
The paper revisits experimental settings for evaluating top-N item recommendation algorithms, focusing on dataset splitting, sampled metrics, and domain selection, and provides findings from extensive experiments on a large dataset to offer suggestions for appropriate setup.
Top-N item recommendation has been a widely studied task from implicit feedback. Although much progress has been made with neural methods, there is increasing concern on appropriate evaluation of recommendation algorithms. In this paper, we revisit alternative experimental settings for evaluating top-N recommendation algorithms, considering three important factors, namely dataset splitting, sampled metrics and domain selection. We select eight representative recommendation algorithms (covering both traditional and neural methods) and construct extensive experiments on a very large dataset. By carefully revisiting different options, we make several important findings on the three factors, which directly provide useful suggestions on how to appropriately set up the experiments for top-N item recommendation.