With the popularity of higher education during recent years, universities and colleges have had more and more researches on assessments to enhance students’ learning performance. Practically performed, however majority schools have limited counselors. In addition, traditional assessment is mostly pen-and-paper tests, therefore, the results are restricted. Take the Learning and Study Strategies Inventory (LASSI) for example, exam participants have to answer questions from more than ten scales of study strategy with 87 assessment items, which is time-consuming and easily resulting in student’s resistance, fatigue and unwillingness to complete the assessment. Therefore, it is difficult to reach expected effect.
To improve foregoing situation, in this dissertation, we come up with a fuzzy data mining technique for the LASSI. Two major steps are taken to do so. First step is to extract valuable or critical questions from questionnaires to directly reduce the number of assessment questions for LASSI, according to the classification charts of decision tree analysis. Second step is to find the related scale of study strategy from association rule analysis to indirectly decrease the correlative scale of study strategy and reduce the assessment questions for LASSI. Moreover, by integrating the concepts of fuzzy set theory, the rules discovered by data mining techniques are assembled as tree structure, and the ways in the past to answer questions from the first to the end are changed with students’ answer results and then the results will be evaluated to decide whether further assessment is required.
A web-based Learning and Study Strategy self-assessment system (Web-LSA) is developed in this dissertation. It is not to replace the original LASSI assessment but to minimize the assessment questions for approximate result. Fewer questions and the web-based system will enhance students’ willingness to more quickly do self-assessment. The results of the self-assessment will be provided counselors to help in finding high-risk students with study disturbances. Therefore, counselors can pay their attention only on these students, which can not only cut down human resource and counseling cost, but make student’s learning performance more efficiently as well. Furthermore, fuzzy data mining techniques can also be applied to social scientific researches, and can be especially efficient and practical in simplifying the assessment questions.