Shahid Rajaee Teacher Training UniversityTechnology of Education Journal (TEJ)2008-04413320090622Adaptive Assessment and Knowledge Level Estimation by Means of Knowledge Model and Bayesian NetworksAdaptive Assessment and Knowledge Level Estimation by Means of Knowledge Model and Bayesian Networks151164132410.22061/tej.2009.1324FAA.KardanAdvanced Technology Laboratory in Electronic Education, Amirkabir University of Technology,Tehran,IranS.KardanFaculty of Computer Engineering and Information Technology, Amirkabir University of Technology,Tehran,IranJournal Article20200527<strong>One of the important issues in a learning management system is the assessment of the learner’s knowledge level. In basic methods of assessment, questions are simply ticked by the learners, and then automatically scored by machine. In this sort of examinations, two problems may be raised. Firstly, the learner may answer and mark the questions accidentally by guessing the answer without having enough knowledge about the related subject. Secondly, he/she may slip in answering the questions due to his/her inattention, although he/she has enough knowledge about the related subject. In this work, an improved method for knowledge level assessment is presented. In the proposed method learning concepts are modeled based on the hierarchical construct of learning objectives, and the learner’s knowledge model is used to estimate his/her knowledge level. This model is established on the basis of Bayesian networks, and considering the hierarchical construct of learning objectives. To evaluate the proposed method, a set of questions by considering the learning objectives was designed and ordered into different levels of complications. Then these questions were used for assessments. In this work also a new algorithm for updating the nodes in a Bayesian network is introduced. By means of this algorithm the effect of guessing and slipping answers in an assessment is undoubtedly reduced. Finally, by presenting a model for knowledge and use of the hierarchical construct of learning objectives, an effective solution for three above mentioned problems of guessing answers, slipping answers, and the origin of wrong answers have been established.</strong><strong>One of the important issues in a learning management system is the assessment of the learner’s knowledge level. In basic methods of assessment, questions are simply ticked by the learners, and then automatically scored by machine. In this sort of examinations, two problems may be raised. Firstly, the learner may answer and mark the questions accidentally by guessing the answer without having enough knowledge about the related subject. Secondly, he/she may slip in answering the questions due to his/her inattention, although he/she has enough knowledge about the related subject. In this work, an improved method for knowledge level assessment is presented. In the proposed method learning concepts are modeled based on the hierarchical construct of learning objectives, and the learner’s knowledge model is used to estimate his/her knowledge level. This model is established on the basis of Bayesian networks, and considering the hierarchical construct of learning objectives. To evaluate the proposed method, a set of questions by considering the learning objectives was designed and ordered into different levels of complications. Then these questions were used for assessments. In this work also a new algorithm for updating the nodes in a Bayesian network is introduced. By means of this algorithm the effect of guessing and slipping answers in an assessment is undoubtedly reduced. Finally, by presenting a model for knowledge and use of the hierarchical construct of learning objectives, an effective solution for three above mentioned problems of guessing answers, slipping answers, and the origin of wrong answers have been established.</strong>https://jte.sru.ac.ir/article_1324_e7e3bc678c53d5e7442efa79d2f11047.pdf