Modeling e-learning system in high education by DoctuS knowledge-based system: the case of Croatia

E-learning is a content management conception which provides learning environment for the students and teachers by delivering digital content. It is becoming increasingly significant to improve the efficiency and effectiveness of teaching and learning and develop interactive and multimedia e-learning modules in university education. Also, e-learning helps to realize the life-long learning and learning society ; it provides location and time-independent education and enables more people participating in higher education. The main problems in implementation of e-learning are: willingness, motivation, quality of educational content on Internet, costs, and training of the designer, the users, and administrator ; how to handle the copyrights, how to assess students ; successes, which technical requirements are necessary to users, and how to realize administrative support. Although the North America will continue to dominate, Asian and European e-learning initiatives are also expanding. In the paper, we will examine the drivers in the e-learning market and discuss the value-added chain of e-learning technologies. Also, we present our comprehension of the constituting elements of the e-learning conception, such as knowledge increase, e-books, knowledge visualization and progress reporting. Main purpose of the paper is to provide understanding of the most important characteristics of the e-learning conception. Using the DoctuS Knowledge-Based System, we propose a model, developed to support the decision-making about the characteristics of the e-learning solution that would fit our system of higher education. DoctuS uses symbolic artificial intelligence that model the way real experts make decisions. There are two basic ways of reasoning in DoctuS. If we use deduction, also called rule-based reasoning, an expert provides the attributes and their values, organizes the attributes into a multi-step graph, and defines the l if...thenr logical rules between the values of the attributes. Cases (decision alternatives) are described by case features, i.e. we choose one values for each attribute for every case. There is another approach: instead of acquiring rules, we can acquire cases of experience, from which DoctuS induces the rules (using a modified ID3 algorithm). This is called induction or case-based reasoning. Experts and decision makers usually formulate more rules than they actually use, for the sake of reliability. With DoctuS, we can reduce the number of rules to those meta-schemata that really affect the decision. These meta-schemata derive from both explicit and tacit knowledge. This third way of reasoning we call reduction. As result we get the same decision using fewer attributes. Our model describes e-learning as a function of four major areas: graphical design, functioning, content and technical issues. Those main areas are described with 26 further attributes. Results of the research would enable us to develop an adequate e-learning system for higher education in Croatia.