Exploring the Space of IR Functions

Abstract : In this paper we propose an approach to discover functions for IR ranking from a space of simple closed-form mathematical func-tions. In general, all IR ranking models are based on two basic variables, namely, term frequency and document frequency. Here a grammar for generating all possible functions is defined which consists of the two above said variables and basic mathematical operations -addition, sub-traction, multiplication, division, logarithm, exponential and square root. The large set of functions generated by this grammar is filtered by check-ing mathematical feasibility and satisfiability to heuristic constraints on IR scoring functions proposed by the community. Obtained candidate functions are tested on various standard IR collections and several sim-ple but highly efficient scoring functions are identified. We show that these newly discovered functions are outperforming other state-of-the-art IR scoring models through extensive experimentation on several IR collections. We also compare the performance of functions satisfying IR constraints to those which do not, and show that the former set of func-tions clearly outperforms the latter one 3 .
Liste complète des métadonnées

Cited literature [20 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01118844
Contributor : Massih-Reza Amini <>
Submitted on : Tuesday, February 24, 2015 - 2:14:05 PM
Last modification on : Saturday, December 15, 2018 - 1:49:49 AM
Document(s) archivé(s) le : Tuesday, May 26, 2015 - 5:40:50 PM

File

ECIR2014.pdf
Files produced by the author(s)

Identifiers

Citation

Parantapa Goswami, Simon Moura, Eric Gaussier, Massih-Reza Amini, Francis Maes. Exploring the Space of IR Functions. 36th European Conference on Information Retrieval, Apr 2014, Amsterdam, Netherlands. pp.372 - 384, ⟨10.1007/978-3-319-06028-6_31⟩. ⟨hal-01118844⟩

Share

Metrics

Record views

271

Files downloads

155