ursw 2013 - ump-st plug-in

61
UMP-ST plug-in: a tool for documenting, maintaining, and evolving probabilistic ontologies Rommel N. Carvalho, Henrique A. da Rocha, and Gilson L. Mendes Brazilian Office of the Comptroller General Marcelo Ladeira, Rafael M. de Souza, and Shou Matsumoto Universidade de Brasília Paper - Uncertainty Reasoning for the Semantic Web URSW - ISWC 10/21/2013 - Sydney, Australia

Upload: rommel-carvalho

Post on 07-May-2015

677 views

Category:

Technology


2 download

DESCRIPTION

Presentation given by Rommel N. Carvalho at the 9th International Workshop on Uncertainty Reasoning for the Semantic Web at the 12th International Semantic Web Conference in October 21, 2013, Sydney, Australia. This was a joint work between the Research and Strategic Information Directorate from Brazil's Office of the Comptroller General and the Department of Computer Science from the University of Brasília. Title: UMP-ST plug-in: a tool for documenting, maintaining, and evolving probabilistic ontologies. Abstract: Although several languages have been proposed for dealing with uncertainty in the Semantic Web (SW), almost no support has been given to ontological engineers on how to create such probabilistic ontologies (PO). This task of modeling POs has proven to be extremely difficult and hard to replicate. This paper presents the first tool in the world to implement a process which guides users in modeling POs, the Uncertainty Modeling Process for Semantic Technologies (UMP-ST). The tool solves three main problems: the complexity in creating POs; the difficulty in maintaining and evolving existing POs; and the lack of a centralized tool for documenting POs. Besides presenting the tool, which is implemented as a plug-in for UnBBayes, this papers also presents how the UMP-ST plug-in could have been used to build the Probabilistic Ontology for Procurement Fraud Detection and Prevention in Brazil, a proof-of-concept use case created as part of a research project at the Brazilian Office of the Comptroller General (CGU).

TRANSCRIPT

Page 1: URSW 2013 - UMP-ST plug-in

UMP-ST plug-in: a tool for documenting, maintaining, and evolving probabilistic

ontologiesRommel N. Carvalho, Henrique A. da Rocha, and Gilson L. Mendes

Brazilian Office of the Comptroller General Marcelo Ladeira, Rafael M. de Souza, and Shou Matsumoto

Universidade de Brasília !

Paper - Uncertainty Reasoning for the Semantic Web URSW - ISWC

10/21/2013 - Sydney, Australia

Page 2: URSW 2013 - UMP-ST plug-in

Agenda

���2

Page 3: URSW 2013 - UMP-ST plug-in

Agenda

Introduction

���2

Page 4: URSW 2013 - UMP-ST plug-in

Agenda

Introduction

UMP-ST

���2

Page 5: URSW 2013 - UMP-ST plug-in

Agenda

Introduction

UMP-ST

UnBBayes Plug-in Architecture

���2

Page 6: URSW 2013 - UMP-ST plug-in

Agenda

Introduction

UMP-ST

UnBBayes Plug-in Architecture

UMP-ST Plug-in Use Case

���2

Page 7: URSW 2013 - UMP-ST plug-in

Agenda

Introduction

UMP-ST

UnBBayes Plug-in Architecture

UMP-ST Plug-in Use Case

Conclusion

���2

Page 8: URSW 2013 - UMP-ST plug-in

Introduction

���3Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 9: URSW 2013 - UMP-ST plug-in

Logic + Uncertainty Big Bang

���4Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 10: URSW 2013 - UMP-ST plug-in

Logic + Uncertainty Big BangIn the last decade there has been a significant increase in formalisms that integrate uncertainty representation into ontology languages:

PR-OWL [5–7],

PR-OWL 2 [4, 3],

OntoBayes [20],

BayesOWL [8],

and probabilistic extensions of SHIF(D) and SHOIN(D) [15]

among others.

���4Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 11: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Ontology

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

Page 12: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

Page 13: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

Page 14: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

Page 15: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

P(isFrontFor|valueOfProcurement = >1M,annualIncome = <10k) = 90%

Page 16: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

P(isFrontFor|valueOfProcurement = >1M,annualIncome = <10k) = 90%

My objective is to define and represent a context model for the interoperability of Sensor

Networks. As my background is not computer science, it's

being a little hard to understand how to put in practice a probabilistic

ontology. PhD student, Wageningen University, The Netherlands

Page 17: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

P(isFrontFor|valueOfProcurement = >1M,annualIncome = <10k) = 90%

This seems a very promising tool, but we need to learn how to best make use of it. When

we try to design using UnBBayes, the questions we

are trying to answer is how do you identify which entities are relevant to the problem and how translate them as variables in your system.

Fusion Engineer, EADS Innovation Works, UK

My objective is to define and represent a context model for the interoperability of Sensor

Networks. As my background is not computer science, it's

being a little hard to understand how to put in practice a probabilistic

ontology. PhD student, Wageningen University, The Netherlands

Page 18: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

P(isFrontFor|valueOfProcurement = >1M,annualIncome = <10k) = 90%

This seems a very promising tool, but we need to learn how to best make use of it. When

we try to design using UnBBayes, the questions we

are trying to answer is how do you identify which entities are relevant to the problem and how translate them as variables in your system.

Fusion Engineer, EADS Innovation Works, UK

My objective is to define and represent a context model for the interoperability of Sensor

Networks. As my background is not computer science, it's

being a little hard to understand how to put in practice a probabilistic

ontology. PhD student, Wageningen University, The Netherlands

I am evaluating PR-OWL as aknowledge representation aswell as reasoning formalism.

I'd like to explore if/how it canbe used for applicationsusing resource devices.

PhD student, University of Texas at Arlington, USA

Page 19: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

P(isFrontFor|valueOfProcurement = >1M,annualIncome = <10k) = 90%

This seems a very promising tool, but we need to learn how to best make use of it. When

we try to design using UnBBayes, the questions we

are trying to answer is how do you identify which entities are relevant to the problem and how translate them as variables in your system.

Fusion Engineer, EADS Innovation Works, UK

My objective is to define and represent a context model for the interoperability of Sensor

Networks. As my background is not computer science, it's

being a little hard to understand how to put in practice a probabilistic

ontology. PhD student, Wageningen University, The Netherlands

I am evaluating PR-OWL as aknowledge representation aswell as reasoning formalism.

I'd like to explore if/how it canbe used for applicationsusing resource devices.

PhD student, University of Texas at Arlington, USA

Why use these variables? Why they are connected in such a way? How do you

choose what type of variable it is?

Fusion Engineer, EADS Innovation Works, UK

Page 20: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

OntologyProbabilistic

���5

A probabilistic ontology is an explicit, formal knowledge representation that expresses knowledge about a domain of application. This includes:

Types of entities that exist in the domain;

Properties of those entities;

Relationships among entities;

Processes and events that happen with those entities;

Statistical regularities that characterize the domain;

Inconclusive, ambiguous, incomplete, unreliable, and dissonant knowledge related to entities of the domain;

Uncertainty about all the above forms of knowledge;

where the term entity refers to any concept (real or fictitious, concrete or abstract) that can be described and reasoned about within the domain of application [5].

Person, Procurement, Enterprise, ...

firstName, lastName, procurementNumber, ...

motherOf, ownerOf, isFrontFor ...

analyzing if requirements are met, choosing better proposal, ...

P(isFrontFor|valueOfProcurement = >1M,annualIncome = <10k) = 90%

This seems a very promising tool, but we need to learn how to best make use of it. When

we try to design using UnBBayes, the questions we

are trying to answer is how do you identify which entities are relevant to the problem and how translate them as variables in your system.

Fusion Engineer, EADS Innovation Works, UK

My objective is to define and represent a context model for the interoperability of Sensor

Networks. As my background is not computer science, it's

being a little hard to understand how to put in practice a probabilistic

ontology. PhD student, Wageningen University, The Netherlands

I am evaluating PR-OWL as aknowledge representation aswell as reasoning formalism.

I'd like to explore if/how it canbe used for applicationsusing resource devices.

PhD student, University of Texas at Arlington, USA

Why use these variables? Why they are connected in such a way? How do you

choose what type of variable it is?

Fusion Engineer, EADS Innovation Works, UK

One thing which might be beyond the scope of this tutorial is a write-up about "Art of Modeling with MEBN". Both narration and

the resultant MEBN help in understanding the problem, but how one reach from a problem description to a MEBN at times is not very clear. ... So when it comes to MEBN, how

one decides about the context nodes, input nodes and resident nodes? Most of the times it might be pretty obvious but

sometimes it is not very clear why certain nodes are modeled as input nodes in a fragment when they could also be modeled as context nodes, etc. Should we follow an object-oriented

approach when identifying important entities or should we think in terms of predicate logic, etc.? As a modeler what

drives our thinking process? Professor, Institute of Business Administration, Pakistan

Page 21: URSW 2013 - UMP-ST plug-in

Our Goal

���6Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 22: URSW 2013 - UMP-ST plug-in

Our GoalUncertainty Modeling Process for Semantic Technologies (UMP-ST)

Describes the main tasks involved in creating probabilistic ontologies.

But it is only a guideline for ontology designers.

���6Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 23: URSW 2013 - UMP-ST plug-in

Our GoalUncertainty Modeling Process for Semantic Technologies (UMP-ST)

Describes the main tasks involved in creating probabilistic ontologies.

But it is only a guideline for ontology designers.

UMP-ST plug-in overcomes three main problems:

the complexity in creating probabilistic ontologies;

the difficulty in maintaining and evolving existing probabilistic ontologies; and

the lack of a centralized tool for documenting probabilistic ontologies.

���6Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 24: URSW 2013 - UMP-ST plug-in

UMP-ST

���7Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 25: URSW 2013 - UMP-ST plug-in

Methodology

���8Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 26: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Page 27: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Page 28: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Page 29: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 30: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 31: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 32: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 33: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 34: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 35: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

Modeling Cycle - Procurement Fraud

���9

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 36: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

UMP-ST Plug-in

���10

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

Page 37: URSW 2013 - UMP-ST plug-in

Introduction - UMP-ST - UnBBayes Plug-in Architecture - UMP-ST Plug-in Use Case - Conclusion

UMP-ST Plug-in

���10

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

“Requirements traceability refers to the ability to describe and follow the

life of a requirement, in both forward and backward

directions.” [11]

Page 38: URSW 2013 - UMP-ST plug-in

UnBBayes Plug-in Architecture

���11Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 39: URSW 2013 - UMP-ST plug-in

UnBBayes Plug-in Framework

���12Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 40: URSW 2013 - UMP-ST plug-in

UnBBayes UMP-ST Plug-in

���13Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 41: URSW 2013 - UMP-ST plug-in

UMP-ST Plug-in Use Case

���14Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 42: URSW 2013 - UMP-ST plug-in

Requirements

���15Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

• 2) Identifique se o comitê organizador de uma licitação deve ser alterado.

– a) Há algum membro do comitê que não tenha ficha limpa?⇤ i) Busque por seu histórico criminal⇤ ii) Busque por suas investigações administrativas

– b) Existe alguma relação entre a comissão e as empresas que participaram delicitações antigas?⇤ i) Procure por um membro da comissão e um responsável da empresa

participante da licitação que estejam relacionados (mãe, pai, irmão ouirmã );

⇤ ii) Procure por um membro da comissão e um responsável da empresaparticipante da licitação que vivam no mesmo endereço.

As figuras 5.1 e 5.2 trazem uma parte da GUI do UMP-ST plugin relativa aos painéisde visualização das hipóteses relacionadas ao dois objetivos em questão.

Figura 5.1: Painel de hipóteses do primeiro objetivo

Figura 5.2: Painel de hipóteses do segundo objetivo

5.3 EntidadesCom os requisitos da ontologia descritos, agora conseguimos analisá-los para começar

a criar a semântica da nossa ontologia. A partir deste passo é necessário registrarmos aorigem de cada elemento criado, me ao backtracking que foi explicado na seção 3.2.2. Este

46

Page 43: URSW 2013 - UMP-ST plug-in

Requirements

���15Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Goal: Find suspicious procurements

Query: Is there any relation between the committee and the enterprises that participated in the procurement?

Evidence: They are siblings

They live at the same address

• 2) Identifique se o comitê organizador de uma licitação deve ser alterado.

– a) Há algum membro do comitê que não tenha ficha limpa?⇤ i) Busque por seu histórico criminal⇤ ii) Busque por suas investigações administrativas

– b) Existe alguma relação entre a comissão e as empresas que participaram delicitações antigas?⇤ i) Procure por um membro da comissão e um responsável da empresa

participante da licitação que estejam relacionados (mãe, pai, irmão ouirmã );

⇤ ii) Procure por um membro da comissão e um responsável da empresaparticipante da licitação que vivam no mesmo endereço.

As figuras 5.1 e 5.2 trazem uma parte da GUI do UMP-ST plugin relativa aos painéisde visualização das hipóteses relacionadas ao dois objetivos em questão.

Figura 5.1: Painel de hipóteses do primeiro objetivo

Figura 5.2: Painel de hipóteses do segundo objetivo

5.3 EntidadesCom os requisitos da ontologia descritos, agora conseguimos analisá-los para começar

a criar a semântica da nossa ontologia. A partir deste passo é necessário registrarmos aorigem de cada elemento criado, me ao backtracking que foi explicado na seção 3.2.2. Este

46

Page 44: URSW 2013 - UMP-ST plug-in

Figura 5.4: Painel de relacionamentos do UMP-ST plugin

podem ser determinísticas ou não determinísticas (que envolvem probabilidade). Abor-darei apenas as regras não determinísticas, uma vez que as regras determinísticas destaontologia estão resumidas a relações de cardinalidade e unicidade.

1. Se um membro do comitê tiver um parente (pai, mãe, irmão ou irmã) responsávelpor descrever os requisitos da licitação, então há mais chances de haver uma relaçãoentre comitê e empresa, o que inibe a concorrência.

2. Se um membro do comitê morar no mesmo endereço do responsável por descreveros requisitos de uma licitação, então há mais chances de haver uma relação entrecomitê e empresa, que diminui a concorrência.

3. Se um contrato de alto valor relacionado a uma licitação tem como responsávelda empresa licitante ganhadora alguém de baixa escolaridade ou que possui umrendimento financeiro anual baixo, então há chances de essa pessoa atuar como“laranja” da firma, o que diminui a competição.

4. Se o responsável da empresa licitante ganhadora também for responsável por outrasempresas que possuem seus CGCs suspensos por participar de outras licitações,então é provável que essa licitação necessite de mais investigação.

5. Se as empresas licitantes são umas relacionadas com as outras, então é provável quea concorrência tenha sido comprometida.

6. Se 1,2,3,4,5, então é provável que a licitação pública precise de mais investigação.

7. Se algum membro do comitê tenha sido condenado por um crime ou tenha sidopenalizado administrativamente, então ele/ela não possui um histórico limpo. Seele/ela foi recentemente investigado, então é provável que ele/ela não tenha umhistórico limpo.

48

Analysis & Design - Entities

���16Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

mecanismo é tratado pelo plugin, que obriga o usuário cadastrar um elemento predecessorno memento de criação da entidade. A ferramenta também permite a geração automáticado backtracking, porém esta não trata palavras omitidas ou que sejam sinônimas ao nomeda entidade. Sendo assim, ainda é necessário que o modelador confirme manualmente queo backtracking esta completo. O quadro de backtracking pode ser visualizado na figura5.6.

As figuras 5.3 e 5.4 foram geradas através do UMP-ST plugin e nos trazem as en-tidades e regras desta ontologia. Uma pessoa tem um nome, um pai e uma mãe (quetambém são pessoas). Todas as pessoas possuem uma identificação única através do seuCPF. A pessoa possui um nível de educação e moraEm um determinado endereço. Cadapessoa emiteDeclaração de seu impostoDeRenda, que inclui o seu faturamentoAnual. OservidorPúblico é uma pessoa que trabalhaPara um órgãoPúblico.

Todas as licitações públicas são requeridas por algum órgãoPúblico, possuem umacomissão formada por um grupo de servidorPúblico e possuem um grupo de empresasparticipantes. Uma dessas será a vencedora da licitação, que receberá um contrato con-tendo o valor do projeto do órgãoPúblico que abriu a licitação. Cada empresa possuipelo menos um representanteLegal e um CGC (numero de cadastro da lista geral de con-tribuintes), que pode ser usado para informar que uma dada empresa esta suspensa departicipar de licitações.Temos também a entidade de investigaçãoAdimistrativa, que teminformações sobre as investigações que envolve um ou mais servidorPúblico. Seu Relató-rio Final, o relatórioJudicialAdministrativo, contém informações sobre a pena aplicada, sehouver. Finalmente temos a entidade investigaçãoCriminal que envolve uma pessoa, como seu Relatório Final, o relatórioJudicialCriminal, que tem informações sobre o veredicto.

Os atributos destas entidades não aparecem nas figuras. Foram criadas 4 atributos:1.Nome (referente a pessoa), 2.Valor (referente ao contrato), 3. estaSuspenso (relativo aCGC) e 4. faturamentoAnual (relativo a impostoDeRenda). Como citado na aberturadeste capítulo, apenas algumas telas serão apresentadas nesta monografia. Caso o leitortenha interesse em ver em detalhes todas as outras entidades, relacionamentos e atributosdesta ontologia há um CD contendo todas as telas.

Figura 5.3: Painel de entidades do UMP-ST plugin

5.4 RegrasAtravés das entidades, com seus atributos e relacionamentos, já criadas conseguimos

criar regras de comportamento da nossa ontologia. Como citado na seção 3.2.3, as regras

47

Page 45: URSW 2013 - UMP-ST plug-in

Figura 5.4: Painel de relacionamentos do UMP-ST plugin

podem ser determinísticas ou não determinísticas (que envolvem probabilidade). Abor-darei apenas as regras não determinísticas, uma vez que as regras determinísticas destaontologia estão resumidas a relações de cardinalidade e unicidade.

1. Se um membro do comitê tiver um parente (pai, mãe, irmão ou irmã) responsávelpor descrever os requisitos da licitação, então há mais chances de haver uma relaçãoentre comitê e empresa, o que inibe a concorrência.

2. Se um membro do comitê morar no mesmo endereço do responsável por descreveros requisitos de uma licitação, então há mais chances de haver uma relação entrecomitê e empresa, que diminui a concorrência.

3. Se um contrato de alto valor relacionado a uma licitação tem como responsávelda empresa licitante ganhadora alguém de baixa escolaridade ou que possui umrendimento financeiro anual baixo, então há chances de essa pessoa atuar como“laranja” da firma, o que diminui a competição.

4. Se o responsável da empresa licitante ganhadora também for responsável por outrasempresas que possuem seus CGCs suspensos por participar de outras licitações,então é provável que essa licitação necessite de mais investigação.

5. Se as empresas licitantes são umas relacionadas com as outras, então é provável quea concorrência tenha sido comprometida.

6. Se 1,2,3,4,5, então é provável que a licitação pública precise de mais investigação.

7. Se algum membro do comitê tenha sido condenado por um crime ou tenha sidopenalizado administrativamente, então ele/ela não possui um histórico limpo. Seele/ela foi recentemente investigado, então é provável que ele/ela não tenha umhistórico limpo.

48

Analysis & Design - Entities

���16Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Person

Procurement

Enterprise

ownerOf

participatesIn

livesAt

mecanismo é tratado pelo plugin, que obriga o usuário cadastrar um elemento predecessorno memento de criação da entidade. A ferramenta também permite a geração automáticado backtracking, porém esta não trata palavras omitidas ou que sejam sinônimas ao nomeda entidade. Sendo assim, ainda é necessário que o modelador confirme manualmente queo backtracking esta completo. O quadro de backtracking pode ser visualizado na figura5.6.

As figuras 5.3 e 5.4 foram geradas através do UMP-ST plugin e nos trazem as en-tidades e regras desta ontologia. Uma pessoa tem um nome, um pai e uma mãe (quetambém são pessoas). Todas as pessoas possuem uma identificação única através do seuCPF. A pessoa possui um nível de educação e moraEm um determinado endereço. Cadapessoa emiteDeclaração de seu impostoDeRenda, que inclui o seu faturamentoAnual. OservidorPúblico é uma pessoa que trabalhaPara um órgãoPúblico.

Todas as licitações públicas são requeridas por algum órgãoPúblico, possuem umacomissão formada por um grupo de servidorPúblico e possuem um grupo de empresasparticipantes. Uma dessas será a vencedora da licitação, que receberá um contrato con-tendo o valor do projeto do órgãoPúblico que abriu a licitação. Cada empresa possuipelo menos um representanteLegal e um CGC (numero de cadastro da lista geral de con-tribuintes), que pode ser usado para informar que uma dada empresa esta suspensa departicipar de licitações.Temos também a entidade de investigaçãoAdimistrativa, que teminformações sobre as investigações que envolve um ou mais servidorPúblico. Seu Relató-rio Final, o relatórioJudicialAdministrativo, contém informações sobre a pena aplicada, sehouver. Finalmente temos a entidade investigaçãoCriminal que envolve uma pessoa, como seu Relatório Final, o relatórioJudicialCriminal, que tem informações sobre o veredicto.

Os atributos destas entidades não aparecem nas figuras. Foram criadas 4 atributos:1.Nome (referente a pessoa), 2.Valor (referente ao contrato), 3. estaSuspenso (relativo aCGC) e 4. faturamentoAnual (relativo a impostoDeRenda). Como citado na aberturadeste capítulo, apenas algumas telas serão apresentadas nesta monografia. Caso o leitortenha interesse em ver em detalhes todas as outras entidades, relacionamentos e atributosdesta ontologia há um CD contendo todas as telas.

Figura 5.3: Painel de entidades do UMP-ST plugin

5.4 RegrasAtravés das entidades, com seus atributos e relacionamentos, já criadas conseguimos

criar regras de comportamento da nossa ontologia. Como citado na seção 3.2.3, as regras

47

Page 46: URSW 2013 - UMP-ST plug-in

Figura5.7:

Exem

plode

telade

ediçãode

regrasdo

UM

P-ST

plugin

52

Analysis & Design - Rules

���17Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Page 47: URSW 2013 - UMP-ST plug-in

Figura5.7:

Exem

plode

telade

ediçãode

regrasdo

UM

P-ST

plugin

52

Analysis & Design - Rules

���17Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

If a member of the committee lives at the same address as a

person responsible for a bidder in the procurement, a relationship is more likely to exist between the committee and the enterprises,

which lowers competition.

Page 48: URSW 2013 - UMP-ST plug-in

Figura5.9:

Exem

plode

telade

ediçãode

gruposdo

UM

P-ST

plugin

54

Analysis & Design - Groups

���18Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 49: URSW 2013 - UMP-ST plug-in

Figura5.9:

Exem

plode

telade

ediçãode

gruposdo

UM

P-ST

plugin

54

Analysis & Design - Groups

���18Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 50: URSW 2013 - UMP-ST plug-in

Analysis & Design - Traceability

���19Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 51: URSW 2013 - UMP-ST plug-in

Conclusion

���20Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 52: URSW 2013 - UMP-ST plug-in

Conclusion

���21Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 53: URSW 2013 - UMP-ST plug-in

ConclusionFirst tool in the world to implement UMP-ST

���21Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 54: URSW 2013 - UMP-ST plug-in

ConclusionFirst tool in the world to implement UMP-ST

Also the first in the world to support the design of POs

���21Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 55: URSW 2013 - UMP-ST plug-in

ConclusionFirst tool in the world to implement UMP-ST

Also the first in the world to support the design of POs

A GUI tool for designing, maintaining, and evolving POs

Overcomes the complexity in creating POs by providing a step by step guidance

Provides a centralized tool for documenting POs

Provides a constant attention to where and what your changes might impact through the implementation of requirements traceability

���21Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 56: URSW 2013 - UMP-ST plug-in

Future Work

���22Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 57: URSW 2013 - UMP-ST plug-in

Future WorkMore tests (still a beta tool)

���22Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 58: URSW 2013 - UMP-ST plug-in

Future WorkMore tests (still a beta tool)

Exporting all documentation to a single PDF of HTML file

���22Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 59: URSW 2013 - UMP-ST plug-in

Future WorkMore tests (still a beta tool)

Exporting all documentation to a single PDF of HTML file

Generating MFrags automatically based on the groups defined in the last step of the Analysis & Design discipline, in order to facilitate the creation of a MEBN model (i.e., PR-OWL PO) during the Implementation discipline

���22Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 60: URSW 2013 - UMP-ST plug-in

Future WorkMore tests (still a beta tool)

Exporting all documentation to a single PDF of HTML file

Generating MFrags automatically based on the groups defined in the last step of the Analysis & Design discipline, in order to facilitate the creation of a MEBN model (i.e., PR-OWL PO) during the Implementation discipline

Apply same methodology to different PO languages

���22Introduction - UMP-ST - UnBBayes Plug-in Architecture -

UMP-ST Plug-in Use Case - Conclusion

Page 61: URSW 2013 - UMP-ST plug-in

Obrigado!

���23