Exploiting Linked Data in Exploratory Search

Pavlos Fafalios

doi:10.1145/3053408.3053427

Abstract

Exploiting Linked Data in Exploratory Search

Highlights

1.1 General ObjectiveWeb Searching is nowadays an integral part of our daily lives
Since the approach that we propose is based on Linked Open Data (LOD), below we first discuss the most relevant LOD-based Named Entity Extraction (NEE) tools of general purpose (3.2.1), we report some semantic annotation systems tailored for the life sciences domain (3.2.2), and we discuss the main differences of our approach (3.2.3)
We distributed the questionnaire to marine biologists and to persons working on marinerelated projects who have a basic knowledge on marine species. 30 subjects participated in the user study (22 to 60 years old), from 6 countries and 12 organizations2

Summary

Introduction

1.1 General ObjectiveWeb Searching is nowadays an integral part of our daily lives. As an indicative example, Google Search Engine [42] processes over 40,000 search queries every second on average which translates to over 3.5 billion searches per day and 1.2 trillion searches per year worldwide. In exploratory or undirected search tasks, where searchers often are unfamiliar with the domain of their goals, unsure about the ways to achieve their goals, and/or need to learn about the topic in order to understand how to achieve their goals (e.g., to find a book, buy a product, perform a patent search, etc.), the classic ranked list of results is not enough [145] In this context, impressions produced out of overview information can play an important role in understanding an information space, and in helping users select promising sub-topics for further exploration. The main objective of this dissertation is to advance exploratory search services by connecting search results with related available semantic information Towards this direction, in this thesis we propose an entity-based exploratory search process that provides searchers with different perspectives of the search results. The outcome of this analysis is exploited for offering add-on services that enhance IR, for grouping the search results (and offering faceted browsing), and for improving them through a re-ranking method

Objectives

Results

Conclusion