A formal study of information retrieval heuristics

By Hui Fang, Tao Tao, Chengxiang Zhai. 2004

375 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems, Information Retrieval & Search

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Systems

In Information Retrieval & Search

Citations in the corpus

A Deep Relevance Matching Model for Ad-hoc Retrieval (2016, background): “In the literature, there have been many formal studies on retrieval models showing the importance of term discrimination [5, 6].”
SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking (2021, neutral): “While it is intuitive that using a log-saturation prevents some terms from dominating – drawing a parallel with axiomatic approaches in IR and log(tf) models [9] – the implied sparsity can seem surprising at first, but, according to our experiments, it obtains better experimental results and allows ”
A Survey of Text Clustering Algorithms (2012, positive): “Effective information retrieval models generally capture three heuristics, i.e., TF weighting, IDF weighting, and document length normalization [36].”
Dense Text Retrieval Based on Pretrained Language Models: A Survey (2022, background): “Another possible approach to understanding the behavior of dense retrievers is axiomatic analysis from classic IR literature [196]–[198].”
Automatic online news issue construction in web environment (2008, methodology): “Through the study of information retrieval heuristics [12], both IWF and IDF satisfy constraints TFC1, TFC2, LNC1 and TF-LNC defined in [12], while IWF behaves more smoothly.”
Learning to model relatedness for news recommendation (2011, neutral): “Among all, BM25 and language models with Dirichlet prior smoothing represent two state-of-theart models in information retrieval [11].”
Semantic term matching in axiomatic approaches to information retrieval (2006, methodology): “The axiomatic approach to information retrieval was proposed recently as a new retrieval framework, in which relevance is modeled by term-based retrieval constraints [5, 6].”
Statistical Language Models for Information Retrieval: A Critical Review (2008, background): “…fusion • Natural language processing for IR • Performance issues for IR systems, including algorithms, data structures, optimization techniques, and scalability • Question answering • Summarization of single documents, multiple documents, and corpora • Text mining • Topic detection and tracking •…”
Improving bug localization using structured information retrieval (2013, positive): “IDF [34], the “probabilistic approach” known as BM25 (Okapi) [32], or more recent language modeling [26], all three approaches have been shown theoretically to utilize the same underlying textual features, and empirically to perform comparably when well-tuned [11].”
Graph-of-word and TW-IDF: new approach to ad hoc IR (2013, methodology): “Regarding improvements that beneﬁt to all these scoring functions, Fang et al. introduced in [6] heuristic retrieval constraints that any weighting model should satisfy.”