Research And Realization Of Adaptive Topic Selection And Automatic Marking

With the development of "Internet + education",online adaptive testing has become a research hotspot.Computer adaptive test(CAT)is a learner centered test,which provides learning paths for learners with different knowledge states.Two problems should be considered in the adaptive test of this thesis: adaptive item selection(adaptive item selection is based on the knowledge state of learners to select an appropriate item for learners)and automatic markng(the marking of learners' answer documents can determine whether or not they master knowledge point,and update the learner's knowledge state in adaptive item selection according to the mastery).In view of the above two problems,the main work of this thesis is as follows:1.The traditional item selection strategy of knowledge space theory and the item selection strategy of solving knowledge state boundary are studied.The knowledge representation of knowledge space theory can be test questions or knowledge points.This thesis first introduces two kinds of application of traditional item selection strategy,traditional item selection strategy based on test questions and traditional item selection strategy based on knowledge points;then introduces two kinds of application of item selection strategy to solve knowledge state boundary,item selection strategy based on knowledge state boundary of test questions and item selection strategy based on knowledge state boundary of knowledge points;compares traditional item selection strategy and solving knowledge state boundary strategy,and analyzes the advantages of the selection strategy of knowledge state boundary;compares the selection strategy of knowledge state boundary based on test questions and analyze the advantages and disadvantages of the both.2.The concept of knowledge block is introduced in the improvement of item selection strategy based on knowledge point and knowledge state boundary.Knowledge block is every divide unit after domain division,is a set of knowledge points.After the division,the scope of item selection is narrowed from one field to each knowledge block.Each knowledge block can select an item in one item selection,so several items selected by all knowledge block are presented to learners as the results of one topic selection.The improved selection process reduces the scope of test knowledge points,simplifies the knowledge structure,shortens the test step length,and has advantages in the flexibility of selection of process assessment range and the distribution regularity of multiple test items in the selection.3.The thesis studies the technology of automatic marking,and puts forward a process from making question to marking.This thesis takes the knowledge points required in the examination of university computer foundation and the operation problems of office software as examples.At the time of making question,the question builder selects the knowledge point and the marking rules of the knowledge point when marking the paper in the making queston interface,and designs description of the test questions,positioning information,standard value,etc.After the operation of making question,the system will generate test questions and parameter files.In the process of marking,the data which is obtained by parsing the answer document in the format of Office Open XML is compared with the standard value of the parameter file to realize automatic marking.In the process of marking papers,the system should deal with the mistakes in the learners' answering documents.4.An adaptive test system is implemented.The main contents include database design,system workflow description,adaptive test demonstration and test results analysis.The rationality and feasibility of the improved adaptive selection process and automatic marking technology are verified.