19,737 Hits in 5.2 sec

Statistical Inference: The Missing Piece of RecSys Experiment Reliability Discourse [article]

Ngozi Ihemelandu, Michael D. Ekstrand
2021 arXiv   pre-print
use of statistical inference in the information retrieval community.  ...  However, there has not yet been significant work on the role and use of statistical inference for analyzing recommender system evaluation results.  ...  use of statistical inference in the information retrieval community (particularly for analyzing TREC search effectiveness metrics).  ... 
arXiv:2109.06424v1 fatcat:ruqnwke4efdxpbbbfalld3vwiy

Exploring Statistical Language Models for Recommender Systems

Daniel Valcarce
2015 Proceedings of the 9th ACM Conference on Recommender Systems  
In this research, we return to the roots of recommender systems and we explore the relationship between Information Filtering and Information Retrieval.  ...  We think that probabilistic methods taken from the latter field such as statistical Language Models can be a more effective and formal way for generating personalised ranks of recommendations.  ...  INTRODUCTION The goal of an Information Retrieval (IR) system is to retrieve the relevant pieces of information according to an information need, typically in the form of a query.  ... 
doi:10.1145/2792838.2796547 fatcat:boysov4tpjcblamiqszbbelgoq

New Metrics for Effective Detection of Shilling Attacks in Recommender Systems

T. Srikanth, Associate Professor, Department of CSE, GITAM University Visakhapatnam, Andhra Pradesh, India, M. Shashi
2019 International Journal of Information Engineering and Electronic Business  
Collaborative filtering techniques are successfully employed in recommender systems to assist users counter the information overload by making accurate personalized recommendations.  ...  However, such systems are shown to be at risk of attacks. Malicious users can deliberately insert biased profiles in favor/disfavor of chosen item(s).  ...  In this paper, two novel metrics are proposed for distinguishing the shilling profiles from normal profiles in collaborative filtering recommender systems.  ... 
doi:10.5815/ijieeb.2019.04.04 fatcat:aayqgqx365cxhfjofxjlkstzta

Towards scalable and accurate item-oriented recommendations

Noam Koenigstein, Yehuda Koren
2013 Proceedings of the 7th ACM conference on Recommender systems - RecSys '13  
Most recommenders research aims at personalized systems, which suggest items based on user profiles. However, in reality many systems deal with item-oriented recommendations.  ...  Second, we address a scalability challenge at the retrieval stage present in many real-world systems: Given an item inventory, which may encompass millions of items, it is desired to identify the most  ...  We quantify the improvement in retrieval time for the task described in (7) The speedup values for each dataset are presented in Table 3 .  ... 
doi:10.1145/2507157.2507208 dblp:conf/recsys/KoenigsteinK13 fatcat:jf37odlj2farvpbhnbdwfs6lue

Performance metrics for the assessment of satellite data products: an ocean color case study

Bridget N. Seegers, Richard P. Stumpf, Blake A. Schaeffer, Keith A. Loftin, P. Jeremy Werdell
2018 Optics Express  
Performance assessment of ocean color satellite data has generally relied on statistical metrics chosen for their common usage and the rationale for selecting certain metrics is infrequently explained.  ...  In contrast, metrics based on simple deviations, such as bias and mean absolute error, as well as pair-wise comparisons, often provide more robust and straightforward quantities for evaluating ocean color  ...  Selection of recommended statistics A variety of statistical performance metrics for algorithm performance assessment exist.  ... 
doi:10.1364/oe.26.007404 pmid:29609296 pmcid:PMC5894891 fatcat:nluxgtxnqnaidpnpa6hdique6u

Quality-biased ranking of web documents

Michael Bendersky, W. Bruce Croft, Yanlei Diao
2011 Proceedings of the fourth ACM international conference on Web search and data mining - WSDM '11  
These content-based features are easy to compute, store and retrieve, even for large web collections.  ...  Table 1 : Summary of feature functions used in a quality-biased sequential dependence model. tfe,D is the number of times e has a match in document D, cfe,D is the number of times concept e matches in  ...  Our quality-biased ranking QSDM outperforms the SDM on all the retrieval metrics (in most cases to a statistically significant degree) for both the unfiltered and the filtered candidate sets.  ... 
doi:10.1145/1935826.1935849 dblp:conf/wsdm/BenderskyCD11 fatcat:px4c5ms4onamvg5eqsowa3vyhm

Fairness in Information Access Systems [article]

Michael D. Ekstrand and Anubrata Das and Robin Burke and Fernando Diaz
2022 arXiv   pre-print
Recommendation, information retrieval, and other information access systems pose unique challenges for investigating and applying the fairness and non-discrimination concepts that have been developed for  ...  We conclude with several open problems in fair information access, along with some suggestions for how to approach research in this space.  ...  background and to lay out consistent terminology for our readers from information retrieval or recommender systems backgrounds.  ... 
arXiv:2105.05779v3 fatcat:fd35qmskibfbfaeblvmiqofgfe

Artist and style exposure bias in collaborative filtering based music recommendations [article]

Andres Ferraro, Dmitry Bogdanov, Xavier Serra, Jason Yoon
2019 arXiv   pre-print
The results of our analysis demonstrate the need for a better evaluation methodology for current music recommendation algorithms, not only limited to user-focused relevance metrics.  ...  In this on-going work we contribute to this research direction analyzing the impact of collaborative filtering recommendations from the perspective of artist and music style exposure given by the system  ...  "Artist and style exposure bias in collaborative filtering based music recommendations", 20th International Society for Music Information Retrieval Conference, Delft, The Netherlands, 2019.  ... 
arXiv:1911.04827v1 fatcat:6wfyc5hozbfaxiqonschbmpzre

Estimation of Fair Ranking Metrics with Incomplete Judgments [article]

Ömer Kırnap, Fernando Diaz, Asia Biega, Michael Ekstrand, Ben Carterette, Emine Yılmaz
2021 arXiv   pre-print
In order to address this problem, we propose a sampling strategy and estimation technique for four fair ranking metrics.  ...  However, the protected attributes of individuals are rarely present, limiting the application of fair ranking metrics in large scale systems.  ...  ACKNOWLEDGMENTS This project was funded by the EPSRC Fellowship titled "Task Based Information Retrieval", grant reference number EP/P024289/1.  ... 
arXiv:2108.05152v1 fatcat:yr6idob6lvdvpk2ptpvp3fx6ie

Improving Accountability in Recommender Systems Research Through Reproducibility [article]

Alejandro Bellogín, Alan Said
2021 arXiv   pre-print
Reasons for this include societal movements around intelligent systems and artificial intelligence striving towards fair and objective use of human behavioral data (as in Machine Learning, Information  ...  In this work, we argue that, by facilitating reproducibility of recommender systems experimentation, we indirectly address the issues of accountability and transparency in recommender systems research  ...  The authors thank the reviewers for their thoughtful comments and suggestions.  ... 
arXiv:2102.00482v1 fatcat:rflwqhxx3fduxfikzdgqza2db4

A Review on Recommender System

L. Anitha, M. Kavitha Devi, P. Anjali Devi
2013 International Journal of Computer Applications  
In this paper we also analyze various issues and evaluation metrics used to measure the performance of the Recommender System.  ...  To address this issue and provide users best recommendations a System is developed called Recommender System.  ...  [1]. 2) Uses keywords for information retrieval.  ... 
doi:10.5120/14098-2115 fatcat:u4l3kzg56rb53otv5jqoem2imq

Validation practices for satellite soil moisture retrievals: What are (the) errors?

A. Gruber, G. De Lannoy, C. Albergel, A. Al-Yaari, L. Brocca, J.-C. Calvet, A. Colliander, M. Cosh, W. Crow, W. Dorigo, C. Draper, M. Hirschi (+14 others)
2020 Remote Sensing of Environment  
We provide theoretical background, a review of state-of-the-art methodologies for estimating errors in soil moisture data sets, practical recommendations on data pre-processing and presentation of statistical  ...  We conclude by identifying research gaps that should be addressed in the near future.  ...  This publication is an outcome of the ISSI's Team on "Adding value to soil moisture information for climate studies" and has received funding from the eartH2Observe project (European Union's Seventh Framework  ... 
doi:10.1016/j.rse.2020.111806 fatcat:wcir35v4drgg7dyezbjtamn6ju

Recommendation or Discrimination?: Quantifying Distribution Parity in Information Retrieval Systems [article]

Rinat Khaziev, Bryce Casavant, Pearce Washabaugh, Amy A. Winecoff, and Matthew Graham
2019 arXiv   pre-print
Information retrieval (IR) systems often leverage query data to suggest relevant items to users.  ...  In this work, we introduce a statistical test for "distribution parity" in the top-K IR results, which assesses whether a given set of recommendations is fair with respect to a specific protected variable  ...  INTRODUCTION Information retrieval (IR) systems, such as search engines and recommender systems (RS), are some of the most widely used machine learning systems today and are used to suggest a list of items  ... 
arXiv:1909.06429v1 fatcat:3mc2b4iyzfdatlz5cx4gkoweuu

Query-Biased Partitioning for Selective Search

Zhuyun Dai, Chenyan Xiong, Jamie Callan
2016 Proceedings of the 25th ACM International on Conference on Information and Knowledge Management - CIKM '16  
A query-biased similarity metric favors terms that are important in query logs. Both methods boost retrieval effectiveness, reduce variance, and produce a more balanced distribution of shard sizes.  ...  This content-based partitioning strategy reveals common topics in a corpus.  ...  Any opinions, findings, conclusions, and recommendations expressed in this paper are the authors' and do not necessarily reflect those of the sponsors.  ... 
doi:10.1145/2983323.2983706 dblp:conf/cikm/DaiXC16 fatcat:vegfhdzrdzcjlaksilho7z7mbu

On the Evaluation of Tweet Timeline Generation Task [chapter]

Walid Magdy, Tamer Elsayed, Maram Hasanain
2016 Lecture Notes in Computer Science  
In this paper, we examine the dependency of TTG on retrieval quality, and its effect on having biased evaluation.  ...  A typical TTG system first retrieves tweets then detects novel tweets among them to form a timeline.  ...  Discussion and Recommendation In this study, we used a set of 13 ad-hoc retrieval runs and 8 TTG systems, resulting in a set of 104 different TTG outputs, which is a reasonable number for getting reliable  ... 
doi:10.1007/978-3-319-30671-1_48 fatcat:75xjqvlwqvh5lblpplyqa36vg4
« Previous Showing results 1 — 15 out of 19,737 results