| Summary: | 碩士 === 亞洲大學 === 資訊工程學系碩士班 === 95 === This research aim inquire in to take Bayesian networks as foundation, combine Knowledge Structure Theory for the principle, Also according to capability beacon, subtechnical ability, false type, create"The Four Basic of Integer" unit of expert Theory and the structure embryo of the Bayesian networks of the Knowledge Structure, and make to answer reaction by student, aim at paper to test sample to carry on a student to try a structure analysis originally, create with the wedge bonding of student's structure and the Bayesian networks three kinds of Bayesian networks models.Purpose understand student at learn the capability beacon of"The Four Basic of Integer" to should it teaching material content, may generate of false type. The setting which also makes use of a Bayesian probability to statistics a method to carry on the analysis of data, and inquire into dissimilarity decision a value influences recognition rate and the Knowledge Structure and student try the influence that a structure down do to know a rate to the whole Bayesian networks at expert, finally create remediable teaching system and identify its result. The results：
1. Combine Bayesian networks a construction to take concept evaluation structure as main evaluation mode by Expert's Knowledge Structure, can be applied in diagnosis student to acknowledge wrong mistake type and sub- technical ability effectively.
2.Want to construct a complete and valid Bayesian networks, need to carry on a cultural heritage study to create first Bayesian networks, make use of again actual data to carry on analysis to revise Bayesian networks.Take Bayesian networks as to inference the test diagnosis of tool system ability valid precise diagnosis, and reach diagnosis result of evaluating the bug or enigma concept that the school-childrens often make.
3.The diagnosis system in this research can complete to test and rectify a teaching activity in the meantime, reaching in the meantime together the result of the work.
4.The computer adaptability of this research development learns system, really canning do an estimate rate stabilization, the adaptability turn and turn separately of teaching purpose.
5.Inference Computerized Adaptive Testing by Bayesian networks, under the situation that the adaptability and the integrity make to answer, the false type, sub- technical ability matches degree to have an effect.
|
|---|
