Estimating Error and Bias in Offline Evaluation Results

January 26, 2020 Β· Declared Dead Β· πŸ› Conference on Human Information Interaction and Retrieval

πŸ‘» CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Mucun Tian, Michael D. Ekstrand arXiv ID 2001.09455 Category cs.IR: Information Retrieval Cross-listed cs.HC Citations 11 Venue Conference on Human Information Interaction and Retrieval Last Checked 4 months ago
Abstract
Offline evaluations of recommender systems attempt to estimate users' satisfaction with recommendations using static data from prior user interactions. These evaluations provide researchers and developers with first approximations of the likely performance of a new system and help weed out bad ideas before presenting them to users. However, offline evaluation cannot accurately assess novel, relevant recommendations, because the most novel items were previously unknown to the user, so they are missing from the historical data and cannot be judged as relevant. We present a simulation study to estimate the error that such missing data causes in commonly-used evaluation metrics in order to assess its prevalence and impact. We find that missing data in the rating or observation process causes the evaluation protocol to systematically mis-estimate metric values, and in some cases erroneously determine that a popularity-based recommender outperforms even a perfect personalized recommender. Substantial breakthroughs in recommendation quality, therefore, will be difficult to assess with existing offline techniques.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Information Retrieval

Died the same way β€” πŸ‘» Ghosted