Generation of Intelligent Assessment Sheet with Compound Criteria in Adaptive Learning System

The computer assisted testing has proven to be an efficient and effective way to evaluate students‟ learning status such that proper tutoring strategies can be adopted to improve their learning performance. In a testing system, the quality of the test items will considerably affect the accuracy of a test; therefore, several measures have been proposed to represent the quality of each test item, such as degree of difficulty and discrimination. However, the quality of a test not only depends on the quality of the item bank, but also relates to the way the assessment sheet is constructed. Selection of appropriate test items is important when constructing a assessment sheet that meets compound assessment requirements, such as expected difficulty degree, expected discrimination degree, number of the test items, and the specified distribution of concept weights. In this paper, a compound criteria assessment sheet generating problem is formulated and an intelligent approach is proposed to cope with the problem. The new approach utilizes the techniques of item response theory, fuzzy logic, enabling construction of a good quality assessment sheet in accordance with specified requirements. Some evaluated results show that the new approach is capable of achieving better performance than previously used test-sheet-generating methods.