Why Design Choices Matter in Recommender Systems

Johannes Kruse; Kasper Lindskow; Michael Riis Andersen; Jes Frellsen

doi:10.1038/s42256-025-01043-5

Why Design Choices Matter in Recommender Systems

Johannes Kruse ^*
, Kasper Lindskow
, Michael Riis Andersen
, Jes Frellsen^*

^*Corresponding author for this work

Department of Management, Society and Communication

Research output: Contribution to journal › Journal article › Research › peer-review

Abstract

In the RecSys ’24 Challenge, participants tackled news recommendations using a large-scale Danish dataset. Although the top-performing models achieved similar accuracy scores, they produced markedly different results on beyond-accuracy metrics, highlighting the need for further research into the normative alignment of recommender systems.

Original language	English
Journal	Nature Machine Intelligence
Volume	7
Issue number	6
Pages (from-to)	979–980
Number of pages	2
ISSN	2522-5839
DOIs	https://doi.org/10.1038/s42256-025-01043-5
Publication status	Published - Jun 2025

Access to Document

10.1038/s42256-025-01043-5Licence: Unspecified

Handle.net

Persistent link

Cite this

@article{691bf76002e34cccaad3a2337d8d42ca,

title = "Why Design Choices Matter in Recommender Systems",

abstract = "In the RecSys {\textquoteright}24 Challenge, participants tackled news recommendations using a large-scale Danish dataset. Although the top-performing models achieved similar accuracy scores, they produced markedly different results on beyond-accuracy metrics, highlighting the need for further research into the normative alignment of recommender systems.",

author = "Johannes Kruse and Kasper Lindskow and \{Riis Andersen\}, Michael and Jes Frellsen",

year = "2025",

month = jun,

doi = "10.1038/s42256-025-01043-5",

language = "English",

volume = "7",

pages = "979–980",

journal = "Nature Machine Intelligence",

issn = "2522-5839",

publisher = "Springer International Publishing",

number = "6",

}

TY - JOUR

T1 - Why Design Choices Matter in Recommender Systems

AU - Kruse , Johannes

AU - Lindskow, Kasper

AU - Riis Andersen, Michael

AU - Frellsen, Jes

PY - 2025/6

Y1 - 2025/6

N2 - In the RecSys ’24 Challenge, participants tackled news recommendations using a large-scale Danish dataset. Although the top-performing models achieved similar accuracy scores, they produced markedly different results on beyond-accuracy metrics, highlighting the need for further research into the normative alignment of recommender systems.

AB - In the RecSys ’24 Challenge, participants tackled news recommendations using a large-scale Danish dataset. Although the top-performing models achieved similar accuracy scores, they produced markedly different results on beyond-accuracy metrics, highlighting the need for further research into the normative alignment of recommender systems.

U2 - 10.1038/s42256-025-01043-5

DO - 10.1038/s42256-025-01043-5

M3 - Journal article

SN - 2522-5839

VL - 7

SP - 979

EP - 980

JO - Nature Machine Intelligence

JF - Nature Machine Intelligence

IS - 6

ER -