An evolutionary strategy with machine learning for learning to rank in information retrieval

Learning to Rank (LTR) is one of the problems in Information Retrieval (IR) that nowadays attracts attention from researchers. The LTR problem refers to ranking the retrieved documents for users in search engines, question answering and product recommendation systems. There is a number of LTR approa...

Full description

Bibliographic Details
Main Authors: Ibrahim, Osman Ali Sadek, Landa-Silva, Dario
Format: Article
Published: Springer 2018
Subjects:
Online Access:https://eprints.nottingham.ac.uk/48943/
_version_ 1848797885432332288
author Ibrahim, Osman Ali Sadek
Landa-Silva, Dario
author_facet Ibrahim, Osman Ali Sadek
Landa-Silva, Dario
author_sort Ibrahim, Osman Ali Sadek
building Nottingham Research Data Repository
collection Online Access
description Learning to Rank (LTR) is one of the problems in Information Retrieval (IR) that nowadays attracts attention from researchers. The LTR problem refers to ranking the retrieved documents for users in search engines, question answering and product recommendation systems. There is a number of LTR approaches based on machine learning and computational intelligence techniques. Most existing LTR methods have limitations, like being too slow or not being very effective or requiring large computer memory to operate. This paper proposes a LTR method that combines a (1+1)-Evolutionary Strategy with machine learning. Three variants of the method are investigated: ES-Rank, IESR-Rank and IESVMRank. They differ on the mechanism to initialize the chromosome for the evolutionary process. ES-Rank simply sets all genes in the initial chromosome to the same value. IESRRank uses linear regression and IESVM-Rank uses support vector machine for the initialization process. Experimental results from comparing the proposed method to fourteen other approaches from the literature show that IESRRank achieves the overall best performance. Ten problem instances are used here, obtained from four datasets: MSLR-WEB10K, LETOR 3 and LETOR 4. Performance is measured at the top-10 query-document pairs retrieved, using five metrics: Mean Average Precision (MAP), Root Mean Square Error (RMSE), Precision (P@10), Reciprocal Rank (RR@10) and Normalized Discounted Cumulative Gain (NDCG@10). The contribution of this paper is an effective and efficient LTR method combining a listwise evolutionary technique with point-wise and pair-wise machine learning techniques.
first_indexed 2025-11-14T20:10:59Z
format Article
id nottingham-48943
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T20:10:59Z
publishDate 2018
publisher Springer
recordtype eprints
repository_type Digital Repository
spelling nottingham-489432020-05-04T19:35:19Z https://eprints.nottingham.ac.uk/48943/ An evolutionary strategy with machine learning for learning to rank in information retrieval Ibrahim, Osman Ali Sadek Landa-Silva, Dario Learning to Rank (LTR) is one of the problems in Information Retrieval (IR) that nowadays attracts attention from researchers. The LTR problem refers to ranking the retrieved documents for users in search engines, question answering and product recommendation systems. There is a number of LTR approaches based on machine learning and computational intelligence techniques. Most existing LTR methods have limitations, like being too slow or not being very effective or requiring large computer memory to operate. This paper proposes a LTR method that combines a (1+1)-Evolutionary Strategy with machine learning. Three variants of the method are investigated: ES-Rank, IESR-Rank and IESVMRank. They differ on the mechanism to initialize the chromosome for the evolutionary process. ES-Rank simply sets all genes in the initial chromosome to the same value. IESRRank uses linear regression and IESVM-Rank uses support vector machine for the initialization process. Experimental results from comparing the proposed method to fourteen other approaches from the literature show that IESRRank achieves the overall best performance. Ten problem instances are used here, obtained from four datasets: MSLR-WEB10K, LETOR 3 and LETOR 4. Performance is measured at the top-10 query-document pairs retrieved, using five metrics: Mean Average Precision (MAP), Root Mean Square Error (RMSE), Precision (P@10), Reciprocal Rank (RR@10) and Normalized Discounted Cumulative Gain (NDCG@10). The contribution of this paper is an effective and efficient LTR method combining a listwise evolutionary technique with point-wise and pair-wise machine learning techniques. Springer 2018-05-01 Article PeerReviewed Ibrahim, Osman Ali Sadek and Landa-Silva, Dario (2018) An evolutionary strategy with machine learning for learning to rank in information retrieval. Soft Computing, 22 (10). pp. 3171-3185. ISSN 1432-7643 Learning to rank ; Evolution strategy ; Linear regression ; Support vector machine https://link.springer.com/article/10.1007%2Fs00500-017-2988-6 doi:10.1007/s00500-017-2988-6 doi:10.1007/s00500-017-2988-6
spellingShingle Learning to rank ; Evolution strategy ; Linear regression ; Support vector machine
Ibrahim, Osman Ali Sadek
Landa-Silva, Dario
An evolutionary strategy with machine learning for learning to rank in information retrieval
title An evolutionary strategy with machine learning for learning to rank in information retrieval
title_full An evolutionary strategy with machine learning for learning to rank in information retrieval
title_fullStr An evolutionary strategy with machine learning for learning to rank in information retrieval
title_full_unstemmed An evolutionary strategy with machine learning for learning to rank in information retrieval
title_short An evolutionary strategy with machine learning for learning to rank in information retrieval
title_sort evolutionary strategy with machine learning for learning to rank in information retrieval
topic Learning to rank ; Evolution strategy ; Linear regression ; Support vector machine
url https://eprints.nottingham.ac.uk/48943/
https://eprints.nottingham.ac.uk/48943/
https://eprints.nottingham.ac.uk/48943/