Improving Results Aggregation Strategies in Distributed Information Retrieval

Author(s):  
Benjamin Ghansah ◽  
Sheng Li Wu ◽  
Nathaniel Ekow Ghansah

The top-ranked documents from various information sources that are merged together into a unified ranked list may cover the same piece of relevant information, and cannot satisfy different user needs. Result diversification(RD) solves this problem by diversifying results to cover more information needs. In recent times, RD has attracted much attention as a means of increasing user satisfaction in general purpose search engines. A myriad of approaches have been proposed in the related works for the diversification problem. However, no concrete study of search result diversification has been done in a Distributed Information Retrieval(DIR) setting. In this paper, we survey, classify and propose a theoretical framework that aims at improving diversification at the result merging phase of a DIR environment.

Author(s):  
Benjamin Ghansah ◽  
Sheng Li Wu

Opposed to centralized search where Websites are crawled and indexed, Distributed Information Retrieval (DIR), also known as Federated Search, is a powerful way to comprehensively search multiple databases in real-time simultaneously. DIR is preferred to centralized search environments in a number of ways, characteristically among them are: 1. the diversity of resources that are made available; 2. improving scalability and reducing server load and network traffic; 3. the leverage of accessing the hidden or deep Web.There are three major phases/tasks of a DIR (i) resource description or collection representation (ii) resource selection and (iii) result merging. This paper aims at providing a comprehensive review on the various phases of DIR and also some current strategies being recommended in enhancing and improving the smooth implementation of a DIR system.


2017 ◽  
Vol 35 (3) ◽  
pp. 398-409
Author(s):  
Gracielle Mendonça Rodrigues Gomes ◽  
Beatriz Valadares Cendon

Purpose The study aims to propose the use of the semiotics inspection method (SIM) which is an interpretative and qualitative method from semiotics engineering (SE) for the evaluation of the communicability of systems and to evaluate digital libraries and information retrieval systems (IRS). The paper presents the results of the application of this method in the evaluation of the quality of the communicability of the interface and search system of the Coordination for the Improvement of Higher Education Personnel (CAPES) Portal of e-Journals, a major scientific digital library in Brazil. There are proposed solutions to improve this system included. Design/methodology/approach The study used the SIM to evaluate the system. Two evaluators inspected the system. They performed the comparison and the analysis of three types of metamessages (metalinguistic, static and dynamic). The metamessages generated by the evaluators were contrasted to find inconsistencies and ambiguities in the CAPES Portal of e-Journals. Finally, the last step of the method was the final assessment about the inspection. Findings The evaluators identified 52 problems of communicability. These problems were ranked according to severity ratings established by Nielsen (1994). They were grouped in ten types of problems present in the interface and in the search system of the CAPES Portal of e-Journals. Originality value This research contributes theoretically to the field of information retrieval and to the area of human–computer interaction and, in particular, to the theory of SE by adapting SE methods that allow the evaluation of communicability to the context of the scientific IRS. Results obtained through scientific methods should contribute to development of the interface and search tools of IRS to better support query formulation and retrieval of relevant information and more efficiently satisfy the information needs of individuals.


Sign in / Sign up

Export Citation Format

Share Document