Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts




НазваниеSupport for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts
страница4/35
Дата05.10.2012
Размер0.88 Mb.
ТипДокументы
1   2   3   4   5   6   7   8   9   ...   35

1.2Aims


The purpose of my research was to test the hypothesis that examining a BN structure by visualisation and verbal explanation of relationships, independently of the numbers, helps in communicating and explaining the model and BN technology to DEs, and thus helps them in constructing a BN of their domains and in defining their understanding of their domains. More specifically, the aims of this research were to:

  1. Propose a rich interactive and integrative approach for exploring the independency assumptions that are represented by the graph topology, using the d-separation criterion in order to support the DE in the BNKE process.




  1. Propose a materialisation of the approach, which includes:

  1. Functionality

  2. A methodology, which is a structured execution of the functionality.




  1. Develop a software tool to implement the functionality ((b), i), which provides the user with:

  1. Interface to choose domain situations to be investigated

  2. Visual and non-technical verbal explanation of the explored situation.




  1. Evaluate the tool. Does examining a BN structure using a tool that includes visualisation and verbal explanation of relationships independently of the numbers, help in communicating and explaining the model and BN technology to DEs, and thus help them in constructing a BN of their domains. This aim can be translated into practical questions regarding the KE process, as follows:




  1. Does the proposed approach help in the assessment of whether the graph correctly represents the domain? In other words, does it help in the investigation of the correctness of the relationships and the influences between the nodes in a network by facilitating the comparison between the structural assumptions of the domain and their graphical modelling?

  2. Does the proposed approach help the DE to gain a better understanding of:

    1. The model.

    2. The domain.

    3. BN technology.

  1. Does the proposed approach support the construction of a BN structure independently of the quantitative part of the model?

  2. Can the proposed approach be utilised for additional tasks?

  3. Can this approach help in reducing the involvement of a BN expert in the process of constructing a network?




  1. Identify limitations of the tool and the approach.



1.3Methods


I designed a visual tool, based on the concepts of the materialisation of the proposed approach, in order to evaluate this approach. The tool, called V-NET, supports the comparison between the structural assumptions of the domain and the graphical modelling decisions. It operates on the graph structure and helps reveal potential misrepresentation of the domain by the graph. The tool has a window interface and uses the structure as the main display. V-NET allows the user to choose nodes and then to choose one of several types of independencies to be displayed. Upon request, V-NET visualises the answer by highlighting the appropriate nodes and displays a verbal explanation of the situation.


Note: There are four different evaluation or assessment processes in this thesis, for which we use the given associated terms:

Evaluating the domain concepts and relationships: Structural assumptions (building ontology).

Evaluating the structural assumptions in order to construct the graph: Graphical modelling (This is based on the structural assumptions but also includes other considerations such as graph complexity, explanatory power, and parameters availability).

Evaluating the “quality” of a constructed graph: Assessment of the graph.

Evaluating the tool: Evaluation.


In order to answer my specific research questions I evaluated the performance of the tool using case studies for qualitative evaluation, employing three different methods of data collection: ‘evaluation’ of the tool using it myself, ‘structured interview’ and ‘unstructured interview’.


These methods were applied in three different settings in order to examine the feasibility of reducing the involvement of a BN expert in the process of KE of a model, and to test different aspects of the KE process.


In the first setting, I assessed so called ‘Toy-Nets’ developed by students as an assignment in a graduate level university course on BNs, as a ‘reality check’. In this setting, I (the BN expert, an ‘outsider’) examined networks designed by others.


In the second setting, a DE analysed his ‘real life’ elicited BN structure that he had previously constructed using V-NET, and answered a structured questionnaire (see Appendix 1) designed by myself (the BN expert). Hence, this setting involved primarily the DE and to some extent, a BN expert.


In the third setting a DE analysed a learnt network of her domain. In this setting the BN expert played only a minimal supportive role, whilst the DE had significant input into the assessment

(section 5.1).
1   2   3   4   5   6   7   8   9   ...   35

Похожие:

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconThis document is being distributed by the American International Health Alliance (aiha) / the EurasiaHealth Knowledge Network with support from the us agency
Саратовский областной Центр по профилактике и борьбе со спид и инфекционными заболеваниями
Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconStudents will demonstrate competence in; Engineering Curriculum; basic concepts such as significant digits, length, and mass; Engineering Drawing; the use of the computer as an engineering tool; designing a project

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconAn Object-Oriented Support Tool for the Design of Casting Procedures

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconRegional network of co-operation in Europe as a tool of efficient governance

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconEnabling Decision Support and Costing of Product Designs by using Visual Metaphors

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconRepresenting and using legal knowledge in integrated decision support systems: DataLex WorkStations

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconThis document contains a list of references to publications and reports about Bayesian Net technology, and especially Bayesian Net applications. The report will

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconNorwegian Network for the Support of the Indigenous Peoples of the Russian Arctic (nnsipra) Сеть Норвежских Организаций в Поддержку Коренных Народов Российского Севера

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconDevelopment of a Product model to support engineering change management

Support for Graphical Modelling in Bayesian Network Knowledge Engineering: a visual Tool for Domain Experts iconTool Use: Misuse of shop tools will result in the loss of tool privileges

Разместите кнопку на своём сайте:
Библиотека


База данных защищена авторским правом ©lib.znate.ru 2014
обратиться к администрации
Библиотека
Главная страница