گزارش هوشمند سیستم آموزش الکترونیکی مبتنی بر وب با استفاده از تکنیک های داده کاوی

This paper presents a PDCA (Plan, Do, Check, Act) method of improving web-based intelligent reports of an e-learning system as intelligent system, which was created and implemented at the Technical Faculty in Cacak, University of Kragujevac. The focus is on improving LMSs (Learning Management Systems) or e-learning systems by predicting behavior patterns of students and adjusting the structure of these electronic courses. An existing learning management system is improved by using data mining techniques and increasing the efficiency of the courses using custom modules. This study presents the design, implementation, and evaluation of the system. Future work should relate to the continued improvement of the PDCA-created system, as well as the introduction of additional modules and a comparative analysis of the presented and future results.

The expansion of e-learning has led to the increased use of systemic and continuous improvement of e-learning systems. This is also evident on the examples of Electrical Engineering teaching courses. The use of these systems has resulted in a need for monitoring and enhancing behavior patterns of all participants, with the aim of continuous improvement of the teaching process and ultimate results – education services. The paper presents the creation of a web-based intelligent report e-learning system using data mining techniques with PDCA (Plan, Do, Check, Act). Learning Management Systems (LMSs), with numerous opportunities in the PDCA, have the ability to track and analyze user activity. Here, administrators can get reports on the activities of participants and statistical approaches at the level of each course (i.e. responsible teacher), as well as at the level of the entire system. Each LMS has a database that contains records on the activities of each user. This characteristic of the system is very significant, with “plenty of information readily available, just a click away” [1, p. 2]. However, many of these records require a special tool for processing and extracting useful information. Such tools have limited capabilities and their use is mostly limited by the administrator’s choice, depending on the type of information. A universal solution for this problem lies in the use of data mining techniques, with the possibility of improving LMS [2] and [3]. Data mining or knowledge discovery in databases (KDDs) is the automatic extraction of implicit and interesting patterns from large data collections [4]. Part of the LMS report also requires improvement in terms of including web intelligence to detect significant patterns of behavior [5]. The need for such a solution includes intelligent and web-based aspects to meet the following requirements (in increments – the PDCA spiral): • The prediction of behavior patterns. • More interactivity. • Visualization of the results obtained. • More real time data analysis. The remainder of this paper has the following structure. Related work is given in Section 2. Section 3 highlights the proposed framework and model, and Section 4 gives the purpose, tasks, and goals of the study. In Section 5, the methodology of creating the system, as well as the creation of dimensions, the OLAP (OnLine Analytical Processing) cube, data mining models, and the architecture of the system is presented. The implementation and evaluation of the system is given in Section 6. Finally, Section 7 highlights the improvement of the system characteristics, as well as its use on the examples of Electrical Engineering course. Concluding remarks are found in Section 8.

A web-based intelligent report e-learning system created with data mining techniques leads to a modern and intelligent way of reporting user activity. Compared with the current reporting system in Moodle LMS, the proposed system presents an improvement since it predicts behavior patterns thus leading to the increased efficiency of the participants. Advantages of the proposed system relate primarily to the opportunities provided by the intelligent features, while a disadvantage is the need for above-average performance of the server. The advantages are reflected in the existing web intelligence and the possibility of predicting user behavior patterns. Furthermore, these advantages have a direct effect on quality assurance in e-learning and on the improvement of the teaching process through the adaptation of content by predicting behavior patterns. Through the Deming PDCA cycle of activities, the following can be concluded (in time t, PtDtCtAt): • Pt: This includes the review of existing reports, and the need for identified reports that provide predictions of user behavior in an LMS. The conclusion at this stage involves moving towards recommendations for planning in terms of the need for inclusion of all relevant factors. • Dt: This includes the design and implementation of the new system and it is consistent with the planned outcomes, adjustment for all activities, and successful functioning of an intelligent, web-based reporting system. • Ct: This is the compulsory phase control, carried out through the testing activities, installation and evaluation systems, the execution of change, and the improvement system. This stage also fulfilled its purpose, and the changes are related to the visualization of results in order to improve them. • At: This includes the possible improvement through the re-use of system analysis tasks. Here, future work should be related to the creation and implementation of new modules in the given time tn, PtnDtnCtnAtn. The approach presented is similar to some previous. However, this study gives improvement in terms of the report system in the field of e-learning. Moreover, development and implementation of new modules, as well as user authentication, will be the subject of future research.