| Abstract: | Abstract
This study attemps to establish the assessment of English tenses to computer adaptive diagnostic test based on Bayesian Network and analyze the error types that students often make in English tenses. According to the error types, apply computer-assisted teaching media and computer remedial teaching media of the establishment to the teaching experiment. The results are as follows.
1. Bayesian network’s error patterns and sub-skills of the average recognition rate was 92.32 %, after carrying on the computerized adaptive test, the average saving rate of test questions reached 16.5%, and reduced testing time. Prediction accuracy on the test questions, the average prediction accuracy was 94.4%, indicating that the prediction accuracy was quite good.
2. In comparative analysis of teaching effectiveness, in the computerized adaptive testing way and traditional teaching way, the achievement of the students of experiment group and control group students has not reached the significant difference.
3. In computer-based remedial instruction, no matter the high-level group, the middle-level group or the lower-level group, the students had good progress, especially in the lower-level group. It reveals computer-based remedial instruction help rase learning effect.
4. Feedback tables based learning found that students thought that computer-assisted teaching and remedial teaching materials and models was helpful to their learning, but when the cognitive respect of difficulty in English tenses, the students thought that the present perfect tense was difficult, only a few people thought that the present progressive tense was difficult, the researchers obtained from these observations considerable feedback, can be used as teaching content and teaching methods in future amendments to the basis.
Key words: Bayesian Networks, tenses, sub-skills, error patterns, Computerized Adaptive Testing |
