Bayesian Network Model for Students' Laboratory Work Performance Assessment: An Empirical Investigation of the Optimal Construction Approach

摘要: There are three approaches to complete Bayesian Network (BN) model construction: total expert-centred, datacentred, and semi data-centred. These constitute the basis of empirical investigation undertaken reported in this paper. The objective is determine, amongst these approaches, which optimal approach for construction a BN-based performance assessment students’ laboratory work virtual electronic environment. BN models were constructed using all with respect focus domain, compared set optimality criteria. In addition, impact size source training, on data-centred was investigated. results provide additional insight constructors contribute literature providing supportive evidence conceptual feasibility efficiency structure parameter learning from data. highlight other interesting themes.