Waste Management Information System

An Expert System Using Ontologies

Erdogan Dogdu, Bahadir Katipoglu and Umutcan Guney

Department of Computer Engineering, TOBB University of Economics and Technology, Ankara, Turkey

Keywords:

Ontology, Expert Systems, Software Design, Knowledge Systems, Waste Management.

Abstract:

Legal documents that include rules and regulations are hard to interpret most of the time. We present the

requirements and the design of an ontology-based expert system framework that we are developing for easy

translation of legal concepts, rules and constraints to an ontology. Ontology engineers and domain experts on

speciﬁc legal documents work collaboratively on this system to generate generic and customized ontologies via

an ontology generation workbench called WOBE (Ontology-based Expert System Workbench) and the expert

system built using the workbench allows end users to follow the rules and regulations without consulting the

complex legal documents. We compare WOBE framework with related tools at the end.

1 INTRODUCTION

Rules and regulations are difﬁcult to follow using

written legal documents. Therefore, domain experts

and consultants are needed in many ﬁelds to help with

compliance with the law and regulations. Waste man-

agement and related environmental laws are one of

those areas that require strict compliance because the

ﬁnes and penalties have been signiﬁcantly increased

in recent years all over the world in the case envi-

ronmental violation of law. We have looked at the

regulations that are overseen by the Waste Manage-

ment Division of The Ministry of Environment and

Urban Planning of Turkey. It consists of 16 separate

documents with over 100.000 words. They are very

detailed and difﬁcult to search and follow. The reg-

ulations are very complex and therefore the Ministry

publishes ﬂow-charts to explain the rules for involved

parties. There are over 700 environmental consulting

companies and more than 11 thousand experts regis-

tered by the Ministry to help companies and institu-

tions with the regulations.

We are developing a research project with fund-

ing from the Ministry of Science, Industry and Tech-

nology of Turkey to develop an expert system that

will help companies, institutions, and individuals to

comply with complex rules and regulations. The sys-

tem will enable domain experts to deﬁne the rules

for any application domain, and end-users to follow

their compliance with the rules. For this purpose an

ontology-based information system will be built with

a dynamic graphical user interface for easy creation

of rules with graphical components in a Web inter-

face. The system will utilize reasoning capabilities

of ontologies to deduce new information and addi-

tional rules, and automatically detect compliance or

non-compliancesituations for the users and alert them

appropriately.

As a case study,we will convert all rules in 16 doc-

uments by the Waste Management Division into the

system with the help of TAYTEK Waste Management

Inc. (Ankara), an environmental consulting company

located in Ankara with domain experts and is a part-

ner and sponsor in our project. We will measure the

effectiveness of the system with test cases in the ﬁeld.

In section 2 we present the motivation for the

project. Then, we present the related work in sec-

tion 3. We present initial prototype design of the sys-

tem in section 4. In the same section we also present

WOBE (Ontology-based Expert System Workbench)

that can be utilized in similar domains as an ontology-

editor and presentation tool; the design and user inter-

face of the WOBE prototype. We present the future

evaluation plans and a comparison of WOBE to simi-

lar tools in section 5. We conclude in section 6.

2 MOTIVATION

We conducted a survey with 58 environmental engi-

neers and environmental consultants from different

regions of Turkey who gathered at a national work-

312

Dogdu E., Katipoglu B. and Guney U..

Waste Management Information System - An Expert System Using Ontologies.

DOI: 10.5220/0005084403120318

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2014), pages 312-318

ISBN: 978-989-758-049-9

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

shop in 2013 in order to discuss the alterations in the

Turkish Environmental Legislation. This survey in-

cluded questions like “Are you making any mistakes

during the determination process of the articles that

will be included in the regulation?”. According to

the results of this survey, 69% of the participants an-

swered “rarely” and 16% answered as “often”. Con-

sequently, 85% of the participants accepted that they

make mistakes in the process at some point. On the

other hand, 72% of the participants agreed that “laws

and regulations are not written in an understandable

language (clearly)”. Among the participants of this

survey, 87% of them were “not aware of expert sys-

tems”, 81% of them were “not aware of decision

trees”, 83% of them were “not aware of ontologies”,

and 91% of them were “not aware of semantic web”.

The results of this survey have shown that regula-

tions are not clear and not easy to follow. And, there

is a need for an expert information system for waste

management that is easy to use, makes rules clear, and

eliminates mistakes. This is also true for many such

regulated domains.

We aim to develop a software system and related

methodologies that can be used to transfer concepts,

relations, and rules in written legal documents to a

knowledgebase. An ontology engineer will use this

system to design an initial generic ontology for a spe-

ciﬁc domain, and a domain expert will customize the

ontology for the speciﬁc decision making processes

in the related ﬁeld of their expertise.

Since any such expert system will need a similar

data and information management infrastructure, we

decided to develop a generic workbench that can be

used to develop similar expert systems. The work-

bench we are developing is called WOBE (Ontology

Based Expert System Workbench) that serves as a

framework and an infrastructure for developing ex-

pert systems.

3 RELATED WORK

There are many works in the literature on expert sys-

tems, ontology-based expert systems, and methods

used in ontology-based expert systems. Here we men-

tion some of the noteworthy work in the area. Ontolo-

gies are now being used in information systems more

frequently than before. We also see more publications

in this area in the literature. Ontology-based informa-

tion systems are used in the domains of law, health-

care, education, science, business, manufacturing and

so on.

There are also many works in the literature on

the design and use of ontologies in legal informa-

tion systems. Kayed presents an early attempt for

an e-law ontology capturing legal concepts and re-

lationships (Kayed, 2005). Gangemi proposes us-

ing Content Ontology Design Patterns which help on-

tology designers to develop legal ontologies easily

(Gangemi, 2007). Khadraoui et al present guidelines

for the development of an eGovernment Information

System ontology towards an eGovernment Informa-

tion System (Khadraoui et al., 2005). (Breuker et al.,

2004), (Cheng et al., 2008), (Lame, 2005), (Wyner

and Hoekstra, 2012), (Wyner, 2008), (van Heijst,

1995)

In JUMAS

(Judicial Management by Digital Lib-

riaries Semantics) project, which is a European Union

supported research project between 2008 and 2011,

tools and methods are developed for semantic enrich-

ment of legal documents (via annotation) for easy dis-

covery and presentation of legal document and multi-

media libraries (audio and video). They developed

a query expansion method and prototype implemen-

tation based on ontologies (Sartori and Palmonari,

2010).

To the best of our knowledge there is no work in

the literature that fully converts a set of legal docu-

ments into a knowledgebase system for testing. In our

project we aim to convert 16 legal documents in waste

management area into a comprehensive ontology and

also develop an expert system that will use these on-

tologies to guide end-users for compliance with those

laws.

Question answering systems are getting popular

with the advances in semantics-based knowledgebase

systems. We also see relevance that in future ex-

pert systems and decision support systems this kind

of semantic question answering will be used more ro-

bustly to guide users towards intelligentdecision mak-

ing using automated reasoning and some sort of nat-

ural language processing. Hakimov et al developed

and tested methods to answer natural language ques-

tions using linked data and relational patterns dis-

covered in the Web (Hakimov et al., 2013). Angele

et al developed an earlier question answering system

that uses semantic Web conceptsand ontologies based

on chemical laws in the context of Digital Aristotle

project (Angele et al., 2003).

In the system we are developing we also deal

with visual presentation and manipulation of ontolo-

gies. In ontology visualization area, Katifori et al sur-

veyed ontology visualization methods (Katifori et al.,

2007). They present the existing methods, evaluate

their characteristics and point to future directions in

ontology visualization. They especially focus on 2D

vs 3D visualization methods and their respective ben-

JUMAS project, http://www.jumasproject.eu

WasteManagementInformationSystem-AnExpertSystemUsingOntologies

313

eﬁts and disadvantages (Katifori et al., 2007). Bosca

et al developed a reusable component for 3D visial-

ization of semantic Web (Bosca et al., 2007). We also

decided to experiment with 3D visualization as a re-

sult of evaluating these work since 3D offer a richer

view and better design approach in ontology visual-

ization and construction.

4 ONTOLOGY-BASED EXPERT

SYSTEM

We are designing an ontology-based expert system

for waste management that will allow ontology engi-

neers, domain experts, and end users to work together.

The backbone of the system is an extended and exten-

sible ontology that will deﬁne the meta-data and rules

in written documents.

Expert system setup process consists of two main

steps: ontology development and ontology usage

phases as depicted in Figure 1. In the ontology devel-

opment phase, ontology engineer creates a “generic

ontology” by translating the artifacts (concepts, re-

lationships, rules) in the legal document to ontology

terms. While creating this ontology,engineer may use

any suitable tool, such as Prot´eg´e

, that is compati-

ble with OWL2 standards. The ontology created at

this phase is a “generic” one because it represents all

of the general concepts and rules in the written legal

document.

GENERIC ONTOLOGY

General Ontology

Instantiation Logic

Instantiation

Develop Use

User Specific Ontology

Ont.Eng./Domain Expert End User

Possible

OWL2

Ontologies

Figure 1: Ontology design and use phases.

In the ontology usage phase the domain experts

and end users customize this generic ontology for

their own needs in user speciﬁc ontologies (Figure 1).

To this end they instantiate the ontological concepts

and relationships by entering domain and/or user spe-

ciﬁc information. In the ontology design phage, on-

tolgy engineer also marks the decision points on the

ontology and therefore decision trees are automati-

cally generated for the domain expert’s use. A sample

decision point, the edge marked S between n

and n

nodes, is shown in Figure 2.

Even though separating the works of ontology en-

gineer and domain expert is not completely possible,

protege.stanford.edu

Figure 2: A sample decision edge.

these decision trees inserted into the ontology isolates

their work to a degree. In the case an ontology engi-

neer needs advice from a domain expert, rather than

communicating with the domain expert, ontology en-

gineer may put decision trees to the necessary points

and let the domain expert make the decision later.

Moreover, the domain expert can complete his or her

design process without any knowledge about ontol-

ogy design. The tool called WOBE, which we are de-

veloping, gives the domain experts a chance of creat-

ing decision trees visually via ﬁnding marked connec-

tions (S edge in the above ﬁgure) in a given ontology.

Apart from this, domain expert or ontology engineer

can analyze the ontology with a three dimensional vi-

sualization if they wish (Figure 7).

Another advantage of inserting the decision trees

into the ontology as described above is that the do-

main expert can add question texts (Figure 8), which

is not in the written text (regulations), to ease the way

to reach to a decision point. Thus, while the general

ontology is a literal translation of the original text, the

decision trees are more user-oriented, asking ques-

tions to ﬁnd out and construct the customized ontol-

ogy instance.

It is now also possible to make inferences on the

acquired ontology by running standard ontology rea-

soning engines on it. These inferences will again be

presented to the user by the client systems. We will

provide this web application within the scope of the

Waste Management Regulations that we are develop-

ing.

4.1 Sample Scenario

Consider the following rule taken from the written

regulation on the Control of Air Pollution Due to In-

dustrial Production of Turkish Environmental Protec-

tion Law:

Article 20b: Air dust emissions can not exceed

the limits set by the regulation for facilities which are

subject to legal authorization for establishment and

operation due to the signiﬁcant impact on air pollu-

tion.

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

314

Concentration limits are:

• CLASS 1 dust emissions (for emission ﬂows of

0.1 kg/hour and above): 20 mg/Nm

• CLASS 2 dust emissions (for emission ﬂows of 1

kg/hour and above): 50 mg/Nm

• CLASS 3 dust emissions (for emission ﬂows of 3

kg/hour and above): 75 mg/Nm

Limits of mixtures:

• CLASS 1, 2: 50 mg/Nm

• CLASS 1, 3 or 2, 3 or 1, 2, 3: 75 mg/Nm

(Class 1,2,3 type materials are listed in the law)

First, an ontology engineer converts this article

into the general ontology terms (Figure 1 develop

phase). Ontology engineer also identiﬁes decision

points in the ontology. For example for the above arti-

cle, dust emission limits can be 20, 50 or 75 mg/Nm3

for class 1, 2, 3 type of materials respectively. We

plan to introduce a visual tool for the engineer where

he or she can choose and mark decision points in the

ontology visually.

Facility

Waste Gas

Emission Limit

hasLimit

[Decision]

20 70 50

Facility

hasLimit

Figure 3: Generic ontology (left) and customized ontology

(right).

In the use phase (Figure 3, 4) the end users de-

cide what to do in decision points that are determined

in the previous stage. And this is done through a

question-answer session on the system. At the end,

the ontology will be customized for the end user with

speciﬁc facts as shown in Figure 3. On the left of the

ﬁgure we see the decision points based on the generic

ontology and on the right side the ﬁnal statement (cus-

tomized ontology) for the end user, that is “the facility

has gas emission limit 70”.

4.2 System Architecture

The expert system we are designing is a multi-tiered

system like all modern information systems. The sys-

tem consists of mainly the following components as

depicted in Figure 5. WOBE, or Ontology Based

Are there any chemicals from CLASS 3 in

emission?

Are there any chemicals from CLASS 2 in

emission?

Are there any chemicals from CLASS 1 in

emission?

yes

Figure 4: A sample decision tree.

Expert System Workbench, is an integrated develop-

ment environment (IDE) for ontology engineers and

domain experts. It serves as a visual tool for these

users at development stage of the expert system build-

ing and it has modules for both ontology creation and

decision tree building.

Expert System Services is a web service that

provides resources for client applications serving

the end user customization stage of the expert sys-

tem. Generic Ontology Framework provides the main

functions for data modelling and customization of on-

tologies. It allows operations like storing, querying,

and updating on ontology model and related decision

trees.

Datastores layer consists of a NoSQL database for

decision trees and a triple store for ontology persis-

tence. Access to this layer is only possible via Generic

Ontology Framework layer.

WOBE

Generic Ontology Framework

Datastores

Expert System Services

Figure 5: Software components of the ontology-based ex-

pert system framework.

4.3 Detailed Architecture

Our initial detailed design is depicted in Figure 6. We

have used some of the open source libraries for spe-

ciﬁc tasks.

Our system uses OWL2

standard and is based on

Jena Framework

for ontology design and manipula-

tion. On top of these, there is another layer we de-

veloped, which we call Generic Ontology Framework

OWL2, www.w3.org/TR/owl2-overview

jena.apache.org

WasteManagementInformationSystem-AnExpertSystemUsingOntologies

315

WOBE: Ontology Based Expert System Workbench

Eclipse 4 RCP

Java 3D JGraphT

Decision Tree Modeller3D Ontology Visualizer

Generic Ontology Framework

Decision Tree DB

Apache Jena Framework

Triple Store

Generic Ontology

Web Service

Client Systems

Figure 6: Software components of the ontology-based ex-

pert system framework.

(GOF). This layer lets the generic ontologies process

using decision trees, save and present them. GOF

stores decision trees in a NoSQL database called De-

cision Tree DB and we use MongoDB

in our proto-

type. We constructed the Generic Ontology Develop-

ment tool named WOBE on Eclipse 4 RCP platform

and used libraries like Java3D

and JGraphT

for 3D

ontology visualization.

4.4 WOBE: Ontology-based Expert

System Workbench

This tool is mainly developed for ontology engineers

and domain experts for designing a “generic ontol-

ogy” which is the base ontology for creating user spe-

ciﬁc ontologies. WOBE is carefully designed to iso-

late the work spaces of ontology engineer and domain

expert from each other. It lets the ontology engineer

create the template ontology with the least knowledge

about the subject and lets the domain expert add the

details having the least knowledge about the ontology.

This is one of the main design principles of WOBE,

that is the “separation of concerns”.

One of the edges in Figure 7 is marked as an edge

which needs a decision model. You can see the de-

sign screen for the decision tree for that edge, which

is used by the domain expert for ontology customiza-

tion in Figure 8.

www.mongodb.org

www.eclipse.org

java3d.java.net

jgrapht.org

Table 1: Comparison table of WOBE with similiar tools

WOBE GEPHI Protege Neoclipse

1.0 0.8.2 4.3 1.9.5

Infrastructure Java Java Java Java

Eclipse 4 Netbeans Eclipse

OWL2 Support

+ - + -

2D Visualization

+ + + +

3D Visualization

+ + +* -

Decision Tree Support

+ - - -

Export to RDF

+ - + -

Graph algorithms

+ + - -

Reasoning

+ + + -

DataStore

TDB Neo4J* TDB Neo4J

∗

with plugin

5 EVALUATION AND

COMPARISON

We are developing WOBE and the expert system soft-

ware. We plan to evaluate the ﬁnished system as fol-

lows:

After creating the waste management regulations

ontology using WOBE, we plan to evaluate the ex-

pert system by user evaluation. For this we employ

our clients and employees of TAYTEK company. We

will conduct a survey about their experience and the

interaction they had with the system after a trial pe-

riod. Thus we will be able to evaluate our systems’

efﬁciency (usability, correctness, etc.). After making

improvements on the system based on the survey re-

sults, we will conduct test with outside users as well.

Here we also present a comparison of features of

WOBE with other ontology design tools (Table 1).

WOBE is not only a graph or ontology development

tool. It is a generic ontology production, editing and

presentation tool that can be used to develop expert

systems for rule-based information management and

decision making.

Gephi is a successful tool on graph visualization

and graph algorithms but it does not support OWL

and RDF standards. Protege is the widely used on-

tology editor but it has no support for decision tree

customization. Neoclipse is an advanced integrated

development environment (IDE) among all.

6 CONCLUSIONS

In this paper we presented the requirements for and

a prototype design of an ontology-based expert sys-

tem that is mainly targeting waste management do-

main. We presented a generic ontology design and

customization tool WOBE and compared its features

with similar tools. Currently we are developing

WOBE. One of the main advantages of WOBE is that

it enables ontology designers and domain experts in

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

316

Figure 7: Prototype view of WOBE.

Figure 8: Prototype view of decision tree designer.

speciﬁc ﬁelds to work together. WOBEs main pur-

pose is to translate concepts, rules, and regulations

written in textdocuments to an information and expert

system via an ontology and related decision trees.

WOBE and its relevant systems are planned to be

licensed by an open-source license. Thus any com-

pany, person or academic foundation can obtain, de-

velop or use freely for building their own expert sys-

tems. The tool will be developed on Eclipse 4 RCP

that has great acceptance from the community and

gives us much more ﬂexibility on extending and de-

veloping.

Companies, which aim to optimize or improve

their processes such as quality control and manage-

ment or systems like CRM and ERM, need computer

assisted autonomous systems. The tool we are devel-

WasteManagementInformationSystem-AnExpertSystemUsingOntologies

317

oping can be used to develop expert systems in any

area by translating rules and regulations to the sys-

tem with the help of an ontology engineer initially.

Then, domain experts can customize the ontology for

speciﬁc usage areas (such as decision making). We

expect to see more sophisticated ontology-based ex-

pert systems in many areas besides legal domain such

as healthcare, business, manufacturing, etc. in near

future.

ACKNOWLEDGEMENT

This work was supported in part by The Ministry of

Science, Industry and Technology of Turkey under

Grant No. SANTEZ-0479.STZ.2013-2.

REFERENCES

Angele, J., M¨onch, E., Oppermann, H., Staab, S., and

Wenke, D. (2003). Ontology-based query and answer-

ing in chemistry: Ontonova project halo. In The Se-

mantic Web-ISWC 2003, pages 913–928. Springer.

Bosca, A., Bonino, D., Comerio, M., Grega, S., and Corno,

F. (2007). A reusable 3d visualization component for

the semantic web. In Proceedings of the twelfth in-

ternational conference on 3D web technology, pages

89–96. ACM.

Breuker, J., Valente, A., and Winkels, R. (2004). Le-

gal ontologies in knowledge engineering and infor-

mation management. Artiﬁcial intelligence and law,

12(4):241–277.

Cheng, C. P., Lau, G. T., Law, K. H., Pan, J., and Jones,

A. (2008). Regulation retrieval using industry spe-

ciﬁc taxonomies. Artiﬁcial Intelligence and Law,

16(3):277–303.

Gangemi, A. (2007). Design patterns for legal ontology

constructions. In LOAIT, pages 65–85.

Hakimov, S., Tunc, H., Akimaliev, M., and Dogdu, E.

(2013). Semantic question answering system over

linked data using relational patterns. In Proceedings of

the Joint EDBT/ICDT 2013 Workshops, pages 83–88.

ACM.

Katifori, A., Halatsis, C., Lepouras, G., Vassilakis, C.,

and Giannopoulou, E. (2007). Ontology visualization

methodsa survey. ACM Computing Surveys (CSUR),

39(4):10.

Kayed, A. (2005). Building e-laws ontology: New ap-

proach. In On the Move to Meaningful Internet Sys-

tems 2005: OTM 2005 Workshops, pages 826–835.

Springer.

Khadraoui, A., Arni-Block, N., L´eonard, M., and Ralyt´e,

J. (2005). Laws-based ontology for e-government in-

formation systems. In The Second international con-

ference on innovations in information technology IIT,

volume 5.

Lame, G. (2005). Using nlp techniques to identify legal

ontology components: concepts and relations. In Law

and the Semantic Web, pages 169–184. Springer.

Sartori, F. and Palmonari, M. (2010). Query expansion for

the legal domain: a case study from the jumas project.

In Ontology, Conceptualization and Epistemology for

Information Systems, Software Engineering and Ser-

vice Science, pages 107–122. Springer.

van Heijst, G. A. C. M. (1995). The role of ontologies in

knowledge engineering.

Wyner, A. (2008). An ontology in owl for legal case-based

reasoning. Artiﬁcial Intelligence and Law, 16(4):361–

387.

Wyner, A. and Hoekstra, R. (2012). A legal case owl ontol-

ogy with an instantiation of popov v. hayashi. Artiﬁ-

cial Intelligence and Law, 20(1):83–107.

KEOD2014-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

318