Keyword-based faceted search interface for knowledge graph construction and exploration

• DOI: https://doi.org/10.1108/IJWIS-02-2022-0037
• Published date: 25 October 2022
• Date: 25 October 2022
• Pages: 453-486
• Subject Matter: Information & knowledge management,Information & communications technology,Information systems,Library & information science,Information behaviour & retrieval,Metadata,Internet
• Author: Samir Sellami,Nacer Eddine Zarour

Keyword-based faceted search

interface for knowledge graph

construction and exploration

Samir Sellami

LIRE Laboratory, Department of Software Technologies and Information Systems,

Abdelhamid Mehri Constantine 2 University Faculty of New Technologies

of Information and Communication, Constantine, Algeria and Department

of Mathematics and Computer Science, Higher Normal School of Technological

Education Skikda, Azzaba, Algeria, and

Nacer Eddine Zarour

LIRE Laboratory, Department of Software Technologies and Information Systems,

Abdelhamid Mehri Constantine 2 University Faculty of New Technologies

of Information and Communication, Constantine, Algeria

Abstract

Purpose –Massive amounts of data, manifesting in various forms, are being produced on the Web

every minute and becoming the new standard. Exploring these information sources distributed in

different Web segments in a unified way is becoming a core task for a variety of users’and companies’

scenarios. However, knowledge creation and exploration from distributed Web data sources is a

challenging task. Several data integration conflicts need to be resolved and the knowledgeneeds to be

visualized in an intuitive manner. The purpose of this paper is to extend the authors’previous

integration works to address semantic knowledge exploration of enterprise data combined with

heterogeneous social and linked Web data sources.

Design/methodology/approach –The authors synthesize information in the form of a knowledgegraph

to resolve interoperability conflicts at integration time. They begin by describing KGMap, a mapping model for

leveraging knowledge graphs to bridge heterogeneous relational, social and linked web data sources. The

mapping model relies on semantic similarity measures to connect the knowledge graph schema with the sources’

metadata elements. Then, based on KGMap, this paper proposes KeyFSI, a keyword-based semantic search

engine. KeyFSI provides a responsive faceted navigating Web user interfaced esignedto facilitate the exploration

and visualization of embedded data behindthe knowledge graph. The authors implemented their approach for a

business enterprise data exploration scenario where inputs are retrieved on the fly from a local customer

relationship management database combined with the DBpedia endpoint and the Facebook Web application

programming interface (API).

Findings –The authors conducted an empirical study to test the effectiveness of their approach using different

similarity measures. The observed results showed better efficiency when using a semantic similarity measure. In

addition, a usability evaluation was conducted to compare KeyFSI features with recent knowledge exploration

systems. The obtained results demonstrate the added value and usability of the contributed approach.

Originality/value –Most state-of-the-art interfaces allow users to browse one Web segment at a time. The

originality of this paper lies in proposing a cost-effective virtual on-demand knowledge creation approach, a

method that enables organizationsto explore valuable knowledge across multiple Web segments simultaneously.

In addition, the responsive components implemented in KeyFSI allow the interface to adequately handle the

uncertainty imposed by thenature of Web information, thereby providing a better user experience.

Keywords Knowledge exploration, Faceted browsing, Keyword search, Responsive interface,

Virtual data exploration, Knowledge graphs, Linked data

Paper type Research paper

Keyword-

based faceted

search

interface

453

Received26 February 2022

Revised17 June 2022

14August 2022

Accepted3 September 2022

InternationalJournal of Web

InformationSystems

Vol.18 No. 5/6, 2022

pp. 453-486

© Emerald Publishing Limited

1744-0084

DOI 10.1108/IJWIS-02-2022-0037

The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1744-0084.htm

1. Introduction

Over the past few decades, the Web has witnessed an unprecedented data avalanche

because of the widespread availability of high-speed internet and the remarkable

progress made in the field of data management. The massive generated data in

different Web segments manifests in various types and formats and rapidly becomes

the new standard on the Web. With the advent of the linked data initiative (Hitzler,

2021), many structured data sources of almost any domain are publicly available. On

the other hand, social media services have grown in popularity, and social data is

becoming more prosperous and important. In many cases, the information scattered

from these various Web segments is complementary and could be connected with

internal enterprise data to find better and meaningful insights. Enterprises need to

access in an integrated way these Web data sources to be able to make their best-

informed business decisions. However, most state-of-the-art interfaces allow users to

browse only one Web segment at a time. Furthermore, information integration and

exploration from distributed Web data sources is a challenging task, as data

providers publish data in various heterogeneous data models. They use different

structures and attribute names to encode the information. Some encode their

information in simple text or legacy relational databases; some produce semi-

structured JSON or XML objects, while others offer machine-readable resource

description framework (RDF) results. To achieve the creation and exploration of

integrated knowledge, several data heterogeneity problems such as technical,

structure or semantics conflicts need to be resolved. Additionally, exploring multiple

Web segments adds a degree of uncertainty to the user interface (UI ), for example,

connectivity problems, longer query execution times and the size of the retrieved

data.

In this context, the recent enterprise knowledge graphs (EKGs) concept is

emerging as a leading solution, capable of mediating between different data models

and effectively supporting the business decision-making process (Hogan et al.,2021).

An EKG represents a linked graph of entities, concepts and facts as well as all the

relationships between them, providing a more precise and complete representation of

an enterprise’sdata(Tiwari et al., 2021). The aim is to connect the concepts and the

elements in the knowledge graph to instance entities in a given collection of

heterogeneous sources to promote data discoverability. The dema nd for knowledge

graphs in enterprises is growing in recent years (Noy et al., 2019). Although the

enterprise domain might vary from marketing to manufacturing, knowledge graphs

can improve performance, reduce costs and establish an additional benefittothe

company’s services (Hislop et al.,2018). For example, Wikidata and DBpedia have

acquired a central position for the Web of Data (Hogan, 2020), while Googl e, acquired

Freebase and transformed it into its own proprietary Knowledge Graph (Nickel et al.,

2016). Building such a knowledge graph for an enterprise requires addressing data

integration challenges to tackle the variety and the increasing number of publicly

available data sets. Data integration aims to synthesize information from diverse data

sources, usually heterogeneous and independent of each other, into a unified

knowledge graph (Villaz

on-Terrazas et al., 2021). Equally important, the data under

the knowledge graph must be easily explored to be actionable, that is, capable of

providing valuable information on the fly to support the decision-making process.

This work extends our paper submitted originally to the KGSWC 2021 conference

(Sellami et al., 2021). In the previous work, we presented a set of complementary tools

for leveraging EKGs by bridging between heterogeneous relational, social and linked

IJWIS

18,5/6

454

data. We suggested MidSemI, a general middleware framework for semantic

integration that resolves interoperability conflicts among heterogeneous sources.

MidSemI defines a knowledge graph as a mediated schema and a metadata model that

allows us to describe source metadata in a flexible way. In this work, we first describe

the KGMap approach, an entity matching approach that exploits the knowledge graph

schema and the different sources’metadata elements to bridge heterogeneous data.

KGMap consists of three sub-models: the knowledge graph representation model,the

source metadata description model and the mapping model. KGMap handles these

models and specifies several possible mapping cases. The goal is to assist an expert to

identify mapping candidates by relying on an algorithm that computes a semantic

similarity measure to determine the relatedness between entities in the metadata of

the source that corresponds to elements in the knowledge graph schema. KGMap is

focused on unifying and integrating heterogeneous data into a knowledge graph.

However, querying remains the primary communication mechanism between the

enterprise knowledge graph and external users or applications. Therefore, based on

the KGMap model, we then propose KeyFSI, a keyword-based faceted search

interface. KeyFSI enables organizations to explore valuable knowledge across

multiple Web segments simultaneously. KeyFSI engine relies on a query-rewriting

algorithm that leverages Natural Language Processing (NLP) techniques to allow

simple keyword queries to be reformulated. The results are then retrieved from the

integrated data synthesized under the knowledge graph. KeyFSI is equipped with a

responsive faceted Web UI that handles adequately the uncertainty imposed by the

nature of Web information, with the goal to provide a better user experience. The

proposed solution implementationis a configurable middleware and a responsive Web

interface capable of exploring internal company data combined on the fly with inputs

from various SPARQL endpoints and Web application programming interfaces

(APIs). Considering an example of a commercial enterpris e, we demonstrate how our

approach leverages enterprise knowledge graph by combining an instance of a

customer relationship management (CRM) database with the SPARQL endpoint of

DBpedia and the Facebook Web API. We conducted an experimental study to test and

compare the effectiveness of KGMap using different similarity me asures. The

observed results suggest that using a semantic-based similarity measure improves

the accuracy of the proposed approach in terms of precision and recall. Through our

use case, we show the value that can be applied to EKGs exploration using KeyFSI.

The implementation of KeyFSI is a multi-faceted navigation interface that uses

modern Web design patterns. We describe the implementation of the KeyFSI UI, then

we conducted a usability evaluation to test its features. The evaluation results

illustrate the feasibility and the ease of use of the proposed approach.

The remainder of this paper is structured as follows: We begin by summarizing

and discussing the limitations of relevant related works in Section2. In Section 3, we

describe a motivating scenario for integrating and exploring embedded knowledge

along with the emerging challenges in this area. Section 4 presents the KGMap model

and the algorithm used to integrate data by calculating direct mappings pairs fro m

heterogeneous internal and external Web sources. The query-rewriting algorithm and

the steps performed by KeyFSI semantic search engine are explained in Section 5.

Section 6 presents the implementation of our approach and provides an empirical

evaluation of our prototype’s accuracy and usability. Finally, Section 7 wraps up and

outlines future work.

Keyword-

based faceted

search

interface

455