Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network
In the early stage of a fire in a building, the human psychological response becomes an important feature in determining survival. Bayesian Network (BN) model has been one of the preferred choices among researchers in representing uncertain knowledge about the psychological responses in the evacuati...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/53853/1/NURULHUDA%20BINTI%20RAMLI%20-%20TESIS%20cut.pdf |
| _version_ | 1797013104817602560 |
|---|---|
| author | Ramli, Nurulhuda |
| author_facet | Ramli, Nurulhuda |
| author_sort | Ramli, Nurulhuda |
| collection | USM |
| description | In the early stage of a fire in a building, the human psychological response becomes an important feature in determining survival. Bayesian Network (BN) model has been one of the preferred choices among researchers in representing uncertain knowledge about the psychological responses in the evacuation system. The BN psychological response models mostly have a literature background of human behaviour in a building on fire. However, the model structure and quantification are almost not involving the opinion of experts in the domain knowledge. Therefore, the goal of this study is to address the uncertainties in the BN model of psychological response during a fire evacuation through expert elicitation. A new conceptual model namely the PRiF (Psychological Response in a Fire) is developed through the social scientific theory and the disaster and fire theory as well as expert opinion approach. For the purpose of model quantification, two phases of data collections involving seven experts from the academic stream and professional fire practitioners are conducted. In the first phase, a new conversion scale for an aided elicitation tool, the Fuzzy Probability Scale is constructed by the experts to elicit the membership functions of the linguistic terms in the scale. The second phase requires the experts to quantify the PRiF model using the Z-number concept, which is able to address the expert’s uncertainty and includes the reliability of their estimates. In doing this, the experts infer their beliefs using the developed conversion scale and provide the confidence of their judgments. |
| first_indexed | 2024-03-06T15:57:21Z |
| format | Thesis |
| id | usm.eprints-53853 |
| institution | Universiti Sains Malaysia |
| language | English |
| last_indexed | 2024-03-06T15:57:21Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | usm.eprints-538532022-08-05T07:18:56Z http://eprints.usm.my/53853/ Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network Ramli, Nurulhuda LC5800-5808 Distance education. In the early stage of a fire in a building, the human psychological response becomes an important feature in determining survival. Bayesian Network (BN) model has been one of the preferred choices among researchers in representing uncertain knowledge about the psychological responses in the evacuation system. The BN psychological response models mostly have a literature background of human behaviour in a building on fire. However, the model structure and quantification are almost not involving the opinion of experts in the domain knowledge. Therefore, the goal of this study is to address the uncertainties in the BN model of psychological response during a fire evacuation through expert elicitation. A new conceptual model namely the PRiF (Psychological Response in a Fire) is developed through the social scientific theory and the disaster and fire theory as well as expert opinion approach. For the purpose of model quantification, two phases of data collections involving seven experts from the academic stream and professional fire practitioners are conducted. In the first phase, a new conversion scale for an aided elicitation tool, the Fuzzy Probability Scale is constructed by the experts to elicit the membership functions of the linguistic terms in the scale. The second phase requires the experts to quantify the PRiF model using the Z-number concept, which is able to address the expert’s uncertainty and includes the reliability of their estimates. In doing this, the experts infer their beliefs using the developed conversion scale and provide the confidence of their judgments. 2020-07 Thesis NonPeerReviewed application/pdf en http://eprints.usm.my/53853/1/NURULHUDA%20BINTI%20RAMLI%20-%20TESIS%20cut.pdf Ramli, Nurulhuda (2020) Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network. PhD thesis, Universiti Sains Malaysia. |
| spellingShingle | LC5800-5808 Distance education. Ramli, Nurulhuda Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network |
| title | Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network |
| title_full | Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network |
| title_fullStr | Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network |
| title_full_unstemmed | Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network |
| title_short | Expert Elicitation Of Uncertainty In Psychological Response In A Fire Evacuation Using Bayesian Network |
| title_sort | expert elicitation of uncertainty in psychological response in a fire evacuation using bayesian network |
| topic | LC5800-5808 Distance education. |
| url | http://eprints.usm.my/53853/1/NURULHUDA%20BINTI%20RAMLI%20-%20TESIS%20cut.pdf |
| work_keys_str_mv | AT ramlinurulhuda expertelicitationofuncertaintyinpsychologicalresponseinafireevacuationusingbayesiannetwork |