DISCOVERING RELATIONSHIP ASSOCIATIONS FROM THE

LITERATURE RELATED TO RESEARCH PROJECTS IN

SUSTAINABILITY SCIENCE USING ONTOLOGY AND

INFERENCE

Weisen Guo and Steven B. Kraines

Division of Project Coordination, the University of Tokyo, Kashiwa-No-Ha, Kashiwa-Shi, Japan

Keywords: Sustainability, Relationship Association, Ontology, Inference, Literature, Research Project.

Abstract: Research projects addressing issues related to sustainability often need knowledge from research papers

from a wide range of disciplines. A method is developed and assessed for using ontology-based inference to

automatically discover knowledge in semantic statements of research papers related to specific research

projects in sustainability science. The semantic statements have been constructed using a semi-automatic

authoring process to represent the knowledge content of the research papers. The discovered knowledge is

expressed in the form of relationship associations that are extracted from semantic statements, where

relationship associations are transitive associations between two binary semantically typed relationships that

share a connecting entity and that co-occur frequently in the set of semantic statements. An algorithm is

presented here for finding interesting relationship associations that are extracted from research papers and

related to a given research project. The method is evaluated on a set of semantic statements containing 104

semantic statements describing research papers and 24 semantic statements describing research projects.

1 INTRODUCTION

The emerging field of sustainability science faces

the challenge of aligning global development with

the resource and waste processing limits of the Earth

in such a way that can be sustained for the

foreseeable future. Many research projects in

sustainability science are being conducted, and these

projects require specialized knowledge from a wide

range of different scientific, sociological, and

economic domains. On the other hand, the rate at

which knowledge resources such as research papers

are published in these different domains is

increasing exponentially. Therefore, how to discover

pre-existing knowledge from different domains that

is both interesting and relevant to particular research

projects in sustainability science is a critical problem.

This paper presents a method to use ontology-

based inference to discover interesting knowledge

from the literature that is related to specific research

projects in sustainability science. The method is

based on “computer understandable” knowledge

descriptors created semi-automatically to describe

the knowledge content of a research paper or project

using terms and relationships provided by an OWL-

DL ontology. The knowledge descriptors, which we

call semantic statements, are computer-

understandable in the sense that a computer can infer

new facts and implications from the descriptors

using logic and rules (Kraines et al., 2006).

The discovered knowledge is expressed in the

form of relationship associations that are extracted

from the descriptors (Guo and Kraines, 2009).

Relationship associations are pairs of binary typed

and directed relationships, often called semantic

triples and occurring in the form “subject – verb ->

object”, that share a connecting entity and that co-

occur frequently in the set of semantic statements.

The method reported here attempts to find the

relationship associations extracted from papers in

the literature that are relevant to a given semantic

statement for a research project. In short, we look for

relationship associations occurring in at least two

research papers, where the first relationship in the

association occurs in the research project semantic

statement. These relationship associations are

hypothesized to be useful for the researchers

involved in the given research project, e.g. to

390

Guo W. and B. Kraines S..

DISCOVERING RELATIONSHIP ASSOCIATIONS FROM THE LITERATURE RELATED TO RESEARCH PROJECTS IN SUSTAINABILITY SCIENCE

USING ONTOLOGY AND INFERENCE.

DOI: 10.5220/0003632603820385

In Proceedings of the International Conference on Knowledge Discovery and Information Retrieval (KDIR-2011), pages 382-385

ISBN: 978-989-8425-79-9

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

identify important related topics that they might

have overlooked. To test the effectiveness of this

knowledge discovery process, we apply it to a set of

semantic statements describing research papers and

research projects related to sustainability science.

The original contribution of this paper is a

method for discovering knowledge in the form of

potentially important new relationships for an entity

that is the focus of a research project from

relationship associations mined from a set of

semantic statements representing scientific papers.

This paper is organized as follows. In Section 2,

we describe our method for discovering related

relationship associations from a set of semantic

statements. In Section 3, we present the results of

applying the method to a set of semantic statements

of research papers and research projects created in

other work. In Section 4, we conclude this paper

with a summary.

2 MATERIALS AND METHODS

Our goal is to discover knowledge in specific

research papers that could be useful to the

researchers involved in particular research projects

in sustainability science.

The first step in our proposed knowledge

discovery method is to mine the set of semantic

statements, which have been created to describe the

knowledge content of individual research papers, for

interesting relationship associations. One

relationship association contains an ordered pair of

semantic triples that share one common entity,

where a semantic triple is comprised of a subject

entity plus a directed and typed relationship plus an

object entity. A relationship association of the form

(e1 –r1-> e2, e1 –r2-> e3) can be interpreted as

follows: if a particular entity e1 has a relationship r1

with another entity e2, then it is likely that e1 has

another relationship r2 with a third entity e3.

For this mining step, we use an algorithm

developed in previous work (Guo and Kraines,

2009; Guo and Kraines, 2010; Guo and Kraines,

2011). Briefly, the algorithm works as follows. The

input of the algorithm is a set of semantic statements,

in this case all of the semantic statements of research

papers. The output of the algorithm is a set of

interesting relationship associations in the form of

linked directed pairs of semantic triples. First, we

extract all of the semantic triples from the set of

semantic statements and convert the triples to triple

queries. Second, we obtain the support for each

triple query in the entire set of semantic statements,

and we discard queries that do not meet a support

threshold. Third, we generate relationship

associations, which express associations between all

pairs of the triple queries that remain. Fourth, we

remove relationship associations that are

semantically equivalent. Fifth, we obtain the support

for each of the remaining relationship associations in

the entire set of semantic statements using logic and

rule based inference.

Next, we want to discover which of the

relationship associations extracted using the mining

process described above are related to each semantic

statement describing a research project. Because we

want relationship associations that are not specific to

a single paper, our first requirement is that the

relationship association must have a support of at

least two, i.e. it must occur in at least two of the

semantic statements for the research papers. Due to

the relatively small number of semantic statements

that we have here (104 statements describing

research papers), we do not use the relevance criteria

from our previous mining algorithm (Guo and

Kraines, 2011). However, if the set of semantic

statements is larger, we can apply those criteria as

well to reduce the size of related and potentially

interesting relationship associations discovered.

We then look for relationship associations where

the first relationship occurs in the semantic

statement for a research project but where the entire

relationship association does not occur. If we find

such a relationship association for a particular

research project, then we can suggest that the

researchers involved in the research project consider

the second relationship as knowledge that they might

want to include in their description of the project.

For example, in one of the projects we are studying,

a researcher is studying an energy system that has a

fuel cell as a component. We have found the

following relationship association from the research

papers: “if a fuel cell is part of an energy system,

then it is likely that the energy system has quantity

efficiency” (hereafter the text in bold font indicates

class and the text in italic font indicates property). If

the project description did not mention the efficiency

of the energy system, we can recommend that the

project researcher consider adding that to the project

description.

3 RESULTS

We evaluate the method using 24 semantic

statements describing research projects that have

been funded by the AGS Promotion Office at the

DISCOVERING RELATIONSHIP ASSOCIATIONS FROM THE LITERATURE RELATED TO RESEARCH

PROJECTS IN SUSTAINABILITY SCIENCE USING ONTOLOGY AND INFERENCE

391

Figure 1: Graph view of the semantic statement for the project entitled “An integrated evaluation system for

countermeasure technologies to realize sustainable societies in cities”. Instances of ontology classes describing entities from

the text abstract are shown with boxes colored according to the major upper class: physical objects are blue, activities are

yellow, events are gray, abstract objects are white. The free text name of the instance is followed by a colon and the name

of the class. Properties specifying the relationship between the instances are shown as directed arrows labeled with the

name of the property.

University of Tokyo between 2000 and 2005 and

104 semantic statements that have been created to

describe research papers related to energy and

sustainability. All 128 of the semantic statements

were created through the EKOSS system (Kraines et

al., 2006) using the SCINTENG ontology (Kraines

and Guo, 2011) as the knowledge representation

language.

From the set of 104 semantic statements

describing research papers, we generated 47,419

semantically unique relationship associations.

However, only 132 relationship associations occur in

the set of semantic statements at least twice. Unlike

concept associations, relationship associations are

directed. Therefore, we can also consider the reverse

of each relationship association, which gives us 264

relationship associations to use in the knowledge

discovery step. Using the algorithm presented in the

previous section, we discovered that, on average, 10

of the 264 relationship associations were related to

each of the 24 research project semantic statements.

One example for the research project entitled

“An integrated evaluation system for

countermeasure technologies to realize sustainable

societies in cities”. The graph view of the semantic

statement created based on the project abstract is

shown in Figure 1. The relationship association that

was discovered to be related to this semantic

statement is:

“If an energy system has part fuel cell, then it is

likely that the energy system has quantity

efficiency.”

This relationship association occurs in the

semantic statements describing the research paper

entitled “Cycle analysis of micro gas turbine-solid

oxide fuel cell hybrid system” (Uechi et al., 2002),

and the research paper entitled “Cycle analysis of

micro gas turbine-molten carbonate fuel cell hybrid

system” (Kimijima and Kasagi, 2005).

Based on this relationship association, we can

suggest that the project researchers might consider

studying the efficiency of the SOFC/GT combined

system, which has a solid oxide fuel cell as a part, as

shown by the triple circled in red in Figure 1.

Another example is for the research project

entitled “development of a strategic traffic model for

the Tokyo urban area and application to the analysis

of control of environmental loading” (We ignore the

graph view of the semantic created based on the

project abstract for the space limitation). The

relationship association that was discovered to be

related to this semantic statement is:

“If a gaseous phase fluid object has material

co2 pollutant, then it is likely that an emission

activity emits that gaseous phase fluid object.”

KDIR 2011 - International Conference on Knowledge Discovery and Information Retrieval

392

This relationship association occurs in the

semantic statements describing the research paper

entitled “Simulation of tradable CO

emission

permits with the New Earth 21 Model” (Yamaji et

al., 1998), and the research paper entitled

“Classifying CO

emissions from the viewpoint of

LCA by reflecting the influence of regional

activities” (Yoshikuni et al., 1998).

Based on this relationship association, we

suggest that the researchers consider that the vehicle

air pollutant emissions result from emission

activities that are sub activities of the commuting

activities.

4 CONCLUSIONS

Sustainability science is fundamentally multi-

disciplinary, and research projects in sustainability

science often need knowledge from many different

research fields. How to effectively utilize the

knowledge existing in the scientific literature to help

researchers resolve issues during their research

projects is a critical problem. We have described and

tested a method to discover knowledge related to

specific research projects in the form of relationship

associations, which are extracted from computer-

understandable descriptors called semantic

statements. We evaluated our method using 104

semantic statements describing research papers and

24 semantic statements describing research projects.

Several relationship associations that we found to be

related to specific research projects appear to

suggest aspects of key entities in the research

projects that might be of interest to the researchers

based on knowledge expressed in the research

papers.

ACKNOWLEDGEMENTS

This work was funded by the Alliance for Global

Sustainability Promotion Office at University of

Tokyo.

REFERENCES

Guo, W., Kraines, S. B., 2009. Discovering Relationship

Associations in Life Sciences Using Ontology and

Inference, In: Proceedings of the 1st International

Conference on Knowledge Discovery and Information

Retrieval 2009, KDIR 2009, pp:10-17, DOI:

10.5220/0002285300100017, SciTePress.

Guo, W., Kraines, S. B., 2010, Mining Relationship

Associations from Knowledge about Failures using

Ontology and Inference, In: Proceedings of the 10th

Industrial Conference on Data Mining ICDM 2010,

Berlin, Germany, July 12-14. In: P. Perner (Ed.):

ICDM 2010, LNAI 6171, pp. 617-631, DOI:

10.1007/978-3-642-14400-4_48, Springer-Verlag

Berlin Heidelberg 2010.

Guo, W., Kraines, S. B., 2011. Extracting Relationship

Associations from Semantic Graphs in Life Sciences,

In: A. Fred et al. (Eds.): IC3K 2010, CCIS 128, pp. 53-

67, 2011, Springer-Verlag Berlin Heidelberg 2011.

Kimijima, S., Kasagi, N., 2005. Cycle analysis of micro

gas turbine-molten carbonate fuel cell hybrid system,

Nippon Kikai Gakkai Ronbunshu, Japan Society of

Mechanical Engineers International Journal Series B,

48(1): 65-74, DOI: 10.1299/jsmeb.48.65.

Kraines, S., Guo, W., Kemper, B., Nakamura, Y., 2006.

EKOSS: A Knowledge-User Centered Approach to

Knowledge Sharing, Discovery, and Integration on the

Semantic Web. In: Cruz, I. et al. (eds) The Semantic

Web – ISWC 2006, LNCS, vol. 4273, pp 833-846. DOI

10.1007/11926078_60.

Kraines, S. B., Guo, W., 2010. Supporting Reuse of

Knowledge of Failures through Ontology-based

Semantic Search, In: Proceedings of the 2nd

International Conference on Knowledge Management

and Information Sharing, KMIS 2010, 25-28 October,

2010, Valencia, Spain, pp: 164-169, DOI:

10.5220/0003068001640169, SciTePress.

Kraines, S. B. and Guo, W., 2011. A system for ontology-

based sharing of expert knowledge in sustainability

science. Data Science Journal, Vol. 9: 107-123, 29

January 2011, DOI: 10.2481/dsj.Kraines.

Uechi, H., Kimijima, S., Kasagi, N., 2002. Cycle analysis

of micro gas turbine-solid oxide fuel cell hybrid

system, Nippon Kikai Gakkai Ronbunshu, B

Hen/Transactions of the Japan Society of Mechanical

Engineers, Part B, 68(666): 626-635.

Yamaji, K., Sato, K., Fujii, Y., Akimoto, K., 1998.

Simulation of tradable CO2 emission permits with the

New Earth 21 Model, International Journal of Global

Energy Issues, 11

Yoshikuni, Y., Hisashi, I., Ryuji, M., 1998. Classifying

CO2 emissions from the viewpoint of LCA by

reflecting the influence of regional activities, Nihon

Enerugi Gakkaishi/Journal of the Japan Institute of

Energy, 77

DISCOVERING RELATIONSHIP ASSOCIATIONS FROM THE LITERATURE RELATED TO RESEARCH

PROJECTS IN SUSTAINABILITY SCIENCE USING ONTOLOGY AND INFERENCE

393