Expert knowledge elicitation for the creation and validation of models in the absence of data

Laura Kreiling, Abigail Hird

Research output: Contribution to conferencePoster

Abstract

In the absence of an abundance of data, resource planners use unstructured estimations as a source of decision making information. Challenges associated with such an approach are: a lack of transparency; slow and lagging response to information demands; poor consistency and agreement; cost intensive collection and ambiguous accuracy (Hird 2012). By developing a technique based on the efficient and structured collection of expert estimations, Hird (2012) overcomes such challenges. In applications to date, a small amount of legacy project data has been available to validate such methods thereby encouraging confidence in model results. This research explores the instance where no data is available for validation of the model. We propose the use of the Delphi method in combination with the technique proposed by Hird (2012). Through case studies, the technique is employed in a UK-based automotive firm. The objectives are: to identify suitable parameters and to evaluate the suitability of Delphi in this context. Initial findings suggest that Delphi is a legitimate means of validating quantitative models and developing confidence in model use. The process of applying Delphi engenders a sense of model ownership and encourages evaluation of current planning practices. Unstructured resource estimation in the absence of data for forecasting model development and validation is a widespread and long standing issue. Our findings also address the issue of expert knowledge retention.

Conference

Conference27th European Conference on Operational Research (EURO XXVII)
CountryUnited Kingdom
CityGlasgow
Period12/07/1515/07/15

Fingerprint

Knowledge acquisition
Transparency
Decision making
Planning
Costs

Cite this

Kreiling, L., & Hird, A. (2015). Expert knowledge elicitation for the creation and validation of models in the absence of data. Poster session presented at 27th European Conference on Operational Research (EURO XXVII), Glasgow, United Kingdom.
Kreiling, Laura ; Hird, Abigail. / Expert knowledge elicitation for the creation and validation of models in the absence of data. Poster session presented at 27th European Conference on Operational Research (EURO XXVII), Glasgow, United Kingdom.
@conference{827227691c624a94a94198b7183473d7,
title = "Expert knowledge elicitation for the creation and validation of models in the absence of data",
abstract = "In the absence of an abundance of data, resource planners use unstructured estimations as a source of decision making information. Challenges associated with such an approach are: a lack of transparency; slow and lagging response to information demands; poor consistency and agreement; cost intensive collection and ambiguous accuracy (Hird 2012). By developing a technique based on the efficient and structured collection of expert estimations, Hird (2012) overcomes such challenges. In applications to date, a small amount of legacy project data has been available to validate such methods thereby encouraging confidence in model results. This research explores the instance where no data is available for validation of the model. We propose the use of the Delphi method in combination with the technique proposed by Hird (2012). Through case studies, the technique is employed in a UK-based automotive firm. The objectives are: to identify suitable parameters and to evaluate the suitability of Delphi in this context. Initial findings suggest that Delphi is a legitimate means of validating quantitative models and developing confidence in model use. The process of applying Delphi engenders a sense of model ownership and encourages evaluation of current planning practices. Unstructured resource estimation in the absence of data for forecasting model development and validation is a widespread and long standing issue. Our findings also address the issue of expert knowledge retention.",
author = "Laura Kreiling and Abigail Hird",
year = "2015",
month = "7",
day = "12",
language = "English",
note = "27th European Conference on Operational Research (EURO XXVII) ; Conference date: 12-07-2015 Through 15-07-2015",

}

Kreiling, L & Hird, A 2015, 'Expert knowledge elicitation for the creation and validation of models in the absence of data' 27th European Conference on Operational Research (EURO XXVII), Glasgow, United Kingdom, 12/07/15 - 15/07/15, .

Expert knowledge elicitation for the creation and validation of models in the absence of data. / Kreiling, Laura; Hird, Abigail.

2015. Poster session presented at 27th European Conference on Operational Research (EURO XXVII), Glasgow, United Kingdom.

Research output: Contribution to conferencePoster

TY - CONF

T1 - Expert knowledge elicitation for the creation and validation of models in the absence of data

AU - Kreiling, Laura

AU - Hird, Abigail

PY - 2015/7/12

Y1 - 2015/7/12

N2 - In the absence of an abundance of data, resource planners use unstructured estimations as a source of decision making information. Challenges associated with such an approach are: a lack of transparency; slow and lagging response to information demands; poor consistency and agreement; cost intensive collection and ambiguous accuracy (Hird 2012). By developing a technique based on the efficient and structured collection of expert estimations, Hird (2012) overcomes such challenges. In applications to date, a small amount of legacy project data has been available to validate such methods thereby encouraging confidence in model results. This research explores the instance where no data is available for validation of the model. We propose the use of the Delphi method in combination with the technique proposed by Hird (2012). Through case studies, the technique is employed in a UK-based automotive firm. The objectives are: to identify suitable parameters and to evaluate the suitability of Delphi in this context. Initial findings suggest that Delphi is a legitimate means of validating quantitative models and developing confidence in model use. The process of applying Delphi engenders a sense of model ownership and encourages evaluation of current planning practices. Unstructured resource estimation in the absence of data for forecasting model development and validation is a widespread and long standing issue. Our findings also address the issue of expert knowledge retention.

AB - In the absence of an abundance of data, resource planners use unstructured estimations as a source of decision making information. Challenges associated with such an approach are: a lack of transparency; slow and lagging response to information demands; poor consistency and agreement; cost intensive collection and ambiguous accuracy (Hird 2012). By developing a technique based on the efficient and structured collection of expert estimations, Hird (2012) overcomes such challenges. In applications to date, a small amount of legacy project data has been available to validate such methods thereby encouraging confidence in model results. This research explores the instance where no data is available for validation of the model. We propose the use of the Delphi method in combination with the technique proposed by Hird (2012). Through case studies, the technique is employed in a UK-based automotive firm. The objectives are: to identify suitable parameters and to evaluate the suitability of Delphi in this context. Initial findings suggest that Delphi is a legitimate means of validating quantitative models and developing confidence in model use. The process of applying Delphi engenders a sense of model ownership and encourages evaluation of current planning practices. Unstructured resource estimation in the absence of data for forecasting model development and validation is a widespread and long standing issue. Our findings also address the issue of expert knowledge retention.

UR - http://euro2015.euro-online.org/

M3 - Poster

ER -

Kreiling L, Hird A. Expert knowledge elicitation for the creation and validation of models in the absence of data. 2015. Poster session presented at 27th European Conference on Operational Research (EURO XXVII), Glasgow, United Kingdom.