Lecture Notes in Computer Science- P90 doc

F. Li et al. (Eds.): ICWL 2008, LNCS 5145, pp. 434 – 445, 2008. © Springer-Verlag Berlin Heidelberg 2008 The Design of Web-Based Personal Collaborative Learning System (WBPCLS) for Computer Science Courses Zhenlong Li 1 and Xiaoming Zhao 1,2 1 Computer Science Department, Taizhou University, Linhai 317000, P.R. China li_zhenlong@163.com 2 Information engineering institute, zhejiang Industrial University, Hangzhou 310014, P.R. China tzxyzxm@yahoo.com.cn Abstract. The Web-based collaborative learning system (WBCLS), which is considered a highly effective teaching method by most theory researchers, could not achieve the goals that can be obtained by traditional in-class teaching method in the teaching practice. Therefore, some problems that are harbored in the operational process of the current popular WBCLS were pointed out. In or- der to overcome problems mentioned above, a new framework for Computer Science Courses, which support the personal learning, was proposed and the al- gorithms about the intelligent course recommendation, optimal group forma- tion, and the optimal collaborative partner discovery are discussed in this paper. Keywords: Web-based Collaborative Learning, Personal Learning, Optimal group formation, Optimal Collaborative partner discovery. 1 Introduction Currently, Web-based learning systems are one of the most interesting topics in the area of the application of computers to education. Using computer technology, espe- cially distributed computing technology, teaching resources share has become reality. Web-based collaborative learning system (WBCLS) , which is considered a highly effective teaching method by most of theory researchers[1, 3], could not achieve the goals which can be obtained by traditional in-class instruction in the teaching prac- tice[2]. The operational process for the WBCLS can be shown in the Fig. 1. According to the operational process mentioned above, we analyzed its reasons for its lower effectiveness than the traditional teaching mode. The reasons are as follow: The first problem for the WBCLS is that improvement of hardware environment has been heavily emphasized, whereas the improvement of learning organization has been heavily ignored in the course of WCBS design. The grouping is one key step of the WBCLS, and is often randomly done without considering the characteristics of individual learners. Therefore, the quality of learning in the collaborative platform is not well achieved as desired. The Design of WBPCLS for Computer Science Courses 435 The second problem is to emphasize excessively the collaborative learning, but ignore the personal learning process. The effectiveness of collaborative learning is different according to the different studying tasks. The effectiveness for some tasks that cannot be partitioned depends on the personal learning process (such as under- standing). Under this situation, it is important to provide appropriate studying re- sources and learning methods for different learners. Learner Login Select learning content Sample grouping Group preparation Group learning Learning evalution Register Learning resources Learner document Learner Characteristic Collaborative strategy Characteristic Extraction Fig. 1. The common operational process for WBCLS The third, the present WBCLS could not intelligently manage the studying record or satisfy learners enough. Because the different courses need different intelligent strategies, the expectation that takes a WBCLS as a common learning environment for all courses is unrealistic. In this paper, the design of the WBPCLS for computer science course will be pre- sented and discussed. The WBPCLS target is to support intelligent grouping and per- sonal learning. 2 The System Design In order to improve the effectiveness of the web-based teaching and overcome the problems in the WBCLS, the operational process of WBPCLS has been designed (Fig. 2) by referencing to the relevant literatures [4-6]. The operational process was summarized in: 1) To login or register: Once a student’s identity is verified by comparison with the corresponding ID in the WBPCLS, the student can enter the WBPCLS environment. The student, who first uses the WBPCLS, needs to fill in some tables for login. The system extracts the student characteristic from the login information; and then builds the personal model database of the student. In the end it adds special measurement mechanisms for measuring the student's characteristic. 436 Z. Li and X. Zhao 2) To select learning content: After the student enters the environment, the system gives student a catalogue of learning content extracted from the studying resource warehouse. Then the student selects one’s learning content. Learner Login Select learning content Intelligent grouping Group preparation Group learning Learning evaluation Register Learning resources Learner document Learner Characteristic Collaborative strategy Characteristic Extraction Learning over? Optimal partner discover Select studying way Need help? Personal learning module Learning over partner learning Group learning yes no no Personal learning yes Fig. 2. The process flowchart for our system 3) To select the learning way: System provides two kinds of learning ways, e.g. the group learning and personal learning. If the student selects the group learning way, steps which students should follow are: ( ) Intelligent group formation. System divides the members into groups and as- signs learning tasks to them based on the studying target, the learner characteristic, and the learning strategy. ( ) The group learning preparation. The learner not only gets acquaintance with other members, environment, and overall target, but also with the roles that the member plays and the collaborative rules. At the same time, the learner extracts the learning target (group and the member target) information from the resource warehouse. ( ) The group learning. The learner studying appointed content, accomplishing the appointed role tasks. The learning resource warehouse provides fundamental material, implements and corresponding function software so that the learning tasks can be completed. System will record interactive behavior in group learning process, stores the information to the corresponding documents, which can be used for the evaluation about member’s learning. The Design of WBPCLS for Computer Science Courses 437 If learner chooses personal learning way , the steps which learners should follow are : ( ) Begin to personal study; ( ) In the learning process, learner, if needs help, can start Optimal Collaborative partner discovery mechanism, which will find out the optimal collaborative partner for current help-seeker; ( ) The collaborative learning with partner. 4)Learning over; 5)Learning evaluation. The WBPCLS modify the corresponding database accord- ing to evaluation result of learner’s behavior in entire learning process, for instance, the learner feature model, the collaborative document. 3 The Learner Feature Selection The learner feature model is the important database, on the basis of which the system is able to provide personal learning and help, realize collaborative learning tasks, and find out the optimal collaborative partner. Learner feature can be acquired by register and characteristic measurement. The register generates the learner’s fundamental information, whereas the characteristic measurement generates the learner’s characteristic warehouse. Both of them consti- tute the learner feature model. For learner's fundamental information, except ID (system assigned), sex, and age, we pay more attention to student's course selecting situation. The system records the courses which the learner has studied, at the same time records student’s score if learner has participated in examination. When learner first enter the collaborative learning system (login), the system re- cords the learner’s selected course that includes course name and course conception. first, Then it measures the learner’s characteristics (Table 1) automatically, and store the result to the corresponding learner’s feature model. The data can be corrected continually according to the learner’s learning situation. Table 1. Learner personal characteristic Character not be very much not Be both will do be Be very much knowledge grade very low low medium high very high collaborative ability very low low medium high very high studying style medium 8~10 field dependent field independent 0~2 2~4 4~6 6~8 We have chosen four characteristics which include the character, the knowledge grade, collaborative ability and studying style to build a learner feature model. The values belong to the range (0~10). 438 Z. Li and X. Zhao System divides "character" into 5 grades: Be ready to help others very much , be ready to help others , both will do not being ready to help others, not being ready to help others very much. The knowledge grade and collaborative ability are also di- vided into 5 grades: high, very high, medium, low, very low. The studying style is divided into field independent tendency or field dependent tendencies. If the value of studying style is in range (4~6), the learner has not obvious field independent and dependent tendency, thus is called "medium ". The knowledge grade was acquired by measuring some knowledge points with grade label, whereas others were measured using the test scale that is composed of an example and 30 tests problem. 4 The Intelligent Course Recommendation In our system, when the learner searches for a course, the system not only can get the relevant information from local resource warehouse or internet, but can also recom- mend the courses according to the learner’s selected courses records. Exampling stu- dent A, we introduce the realization of course recommendation. 1. Course classification In order to reduce course matrix size, we classify the computer science courses using the method of conception extraction. The classification result of all computer science courses and their corresponding conceptions are shown in the Table 2. Table 2. Classification of computer science courses Course name Course conception Network Programming Network Management& Security Distributing System Software Engineering Systems Analysis & Design Multimedia System Design Database Management E-Commerce Programming Management Information System Computer Architecture Multimedia Introduction Data Structure Database Management System Introduction Artificial Intelligence {network,software,method} {network,hardware,management } {network,system,management} {software,method,management} {software,database,method,system} {software,system} {network,database,management} {software,database,method,system,management} { software,system} { software,database,method,system,management} {hardware,method } {software,system } {software,system} { Network, hardware, software, database,method,system,management } { software,method } 2. The course conception favor mining Once student's records of selected courses were taken out from the learning database, the course conception favor can be found by using the Apriori algorithm. For example, . Department, Taizhou University, Linhai 317000, P.R. China li_zhenlong@163.com 2 Information engineering institute, zhejiang Industrial University, Hangzhou 310014, P.R. China tzxyzxm@yahoo.com.cn. are one of the most interesting topics in the area of the application of computers to education. Using computer technology, espe- cially distributed computing technology, teaching resources share. the studying resource warehouse. Then the student selects one’s learning content. Learner Login Select learning content Intelligent grouping Group preparation Group learning Learning evaluation Register Learning