Gwo-Haur Hwang

Diagnosing student learning problems based on historical assessment records

In the past decade, researchers have attempted to develop computer‐assisted learning and testing systems to help students improve their learning performance. Conventional testing systems simply provide students with a score, and do not offer sufficient information in order to improve their learning performance. It would be of more benefit to students if the test results could be critically analysed and hence learning suggestions could be offered accordingly. This study proposes an algorithm for diagnosing students’ learning problems and provides personalised learning suggestions for Science and Mathematics courses. An intelligent tutoring, evaluation and diagnosis system has been implemented based on the novel approach. Experimental results on a Mathematics course have demonstrated the feasibility of this approach in enhancing students’ learning performance, making it highly promising for further study.

A time scale-oriented approach for building medical expert systems☆

Abstract Knowledge acquisition is known to be a critical bottleneck in building expert systems. In past decades, various methods and systems have been proposed to efficiently elicit expertise from domain experts. However, in building a medical expert system, disease symptoms are usually treated as time-irrelevant attributes, such that much important information is abandoned and hence, the performance of the constructed expert systems is significantly affected. To cope with this problem, in this paper, we propose a time scale-oriented approach to eliciting medical knowledge from domain experts. The novel approach takes the time scale into consideration, such that the variant of disease symptoms in different time scales can be precisely expressed. An application to the development of a medical expert system has depicted the superiority of our approach.

A Computer-Assisted Approach to Conducting Cooperative Learning Process

Cooperative learning has been proven to be helpful in enhancing the learning performance of students. The goal of a cooperative learning group is to maximize all members’ learning, which is accomplished via promoting each other’s success, through assisting, sharing, mentoring, explaining, and encouragement. To achieve the goal of cooperative learning, it is very important to organize well-structured cooperative learning groups, in which all group members have the ability to help each other during the learning process. In this article, a concept-based approach is proposed to organize cooperative learning groups, such that, for a given course each concept is precisely understood by at least one of the students in each group. An experiment on a computer science course has been conducted in order to evaluate the efficacy of this new approach. From the experimental results, we conclude that the novel approach is helpful in enhancing student learning efficacy.

A Computerized Approach to Diagnosing Student Learning Problems in Health Education

With the rapid progress of computer technology during recent years, researchers have attempted to develop more effective programs for testing and improving student learning performance. However, in conventional testing systems, students merely obtain a score based on their test results, and are given no direction regarding how to improve their learning performance. Students would benefit from ways of analyzing test results and being provided with learning suggestions. This investigation presents a learning diagnosis approach for providing students with personalized learning suggestions by analyzing their test results and test item related concepts. Based on this approach, a testing and diagnosis system is implemented on computer networks. Experimental results on a nutrition course have demonstrated the feasibility of this approach in enhancing students in their learning performance, making it highly promising for further study.

Development of a Web-Based System for Diagnosing Student Learning Problems on English Tenses

In the past decades, English learning has received lots of attention all over the world, especially for those who are not native English speakers. Various English learning and testing systems have been developed on the Internet. Nevertheless, most existing English testing systems represent the learning status of a student by assigning that student with a score or grade. This approach makes the student aware of his/her learning status through the score or grade, but the student might be unable to improve his/her learning status without further guidance. In this paper, an intelligent English tense learning and diagnostic system is proposed, which is able to identify student learning problems on English verb tenses and to provide personalized learning suggestions in accordance with each student’s learning portfolio. Experimental results on hundreds of college students have depicted the superiority of the novel approach. 1825 Copyright

Design of pseudoexhaustive testable PLA with low overhead

The pseudoexhaustive testing (PET) scheme is an economical approach to testing a large embedded programmable logic array (PLA). The authors propose an efficient algorithm named low overhead PET (LOPET) to partition the product lines. By applying this algorithm, both the area overhead and test length are reduced significantly. >

Analyzing domain expertise by considering variants of knowledge in multiple time scales

Knowledge acquisition is known to be a critical bottleneck in building expert systems. In past decades, various methods and systems have been proposed to efficiently elicit expertise from domain experts. However, in building a medical expert system, disease symptoms are usually treated as time-irrelevant attributes, such that much important information is abandoned and hence the performance of the constructed expert systems is significantly affected. To cope with this problem, in this paper, we propose a time scale-oriented approach to eliciting medical knowledge from domain experts. The novel approach takes the time scale into consideration, such that the variant of disease symptoms in different time scales can be precisely expressed. An application to the development of a medical expert system has depicted the superiority of our approach.

Restructuring and logic minimization for testable PLA

The untestability cube-number product (UCP), a testability measure that can accurately indicate the extra logic needed in testable programmable logic arrays (PLAs), is discussed. Two UCP-based PLA synthesis algorithms are developed. The first one is a restructuring algorithm named REST, and the other is a logic minimizer for testable PLA named LMTPLA. REST can make the restructured PLA testable by taking less extra hardware. LMTPLA is based on EXPRESSO-II and REST. It can consider the testability at the logic minimization process. In order to minimize UCP as well as the number of product terms, four strategies are developed: deleting the cubes with poor testability and reserving the cubes with good testability; giving up the primes, if necessary: partitioning the more untestable cubes into smaller cubes; and deleting the procedures which are useless in LMTPLA. REST and LMTPLA have been implemented on SUN4/260 in C language. For 40 benchmark circuits, the hardware overheads required are reduced by about 30-40%. >

Applying Cluster Techniques of Data Mining to Analysis the Game-Based Digital Learning Work

Clustering is the most important task in unsupervised learning and applications is a major issue in cluster analysis. Digital learning, which arises in recent years, has become a trend of learning method in the future. The environment of digital learning may enable the learners work anytime and everywhere without the limitation of time and space. Another great improvement of digital learning is the ability of recording complete portfolio. These portfolios may be used to gain critical factors of learning if they are analyzed by data mining methods. Therefore, in this research will to analyze the records of students’ portfolios of game-based homework by using Clustering Algorithm Based on Histogram Threshold (HTCA) method of data mining. The HTCA method combines a hierarchical clustering method and Otsu’s method. The result indicates that the attributes or categories of impacting factors and to find conclusions of efficiency for the learning process.

Knowledge verification for fuzzy expert systems

Abstract The introduction and use of fuzzy logic has strengthened knowledge representation and reasoning capability in expert systems; nevertheless, it also increases the complexity and difficulty of knowledge verification, which is known to be an important issue for building reliable and high performance expert systems. In the past decade, knowledge verification problems, e.g., redundancy, conflict, circularity and incompleteness of knowledge, have been widely discussed from the viewpoint of using binary logic; nevertheless, the issue of verifying fuzzy knowledge is seldom discussed. In this paper, we attempt to detect potential structural errors among fuzzy rules by proposing a fuzzy verification algorithm. Moreover, a system for verifying fuzzy knowledge base has been developed based on the novel approach.