در استخراج دانش توصیه برای یادگیری شخصی وب مبتنی بر بهینه سازی کلونی مورچه ها با استراتژی های سگمنتال هدف و کنترل متا

Personalized web-based learning has become an important learning form in the 21st century. To recommend appropriate online materials for a certain learner, several characteristics of the learner, such as his/her learning style, learning modality, cognitive style and competency, need to be considered. An earlier research result showed that a fuzzy knowledge extraction model can be established to extract personalized recommendation knowledge by discovering effective learning paths from past learning experiences through an ant colony optimization model. Though that results revealed the theoretical potential of the proposed method in discovering effective learning paths for learners, critical limitations arose when considering its applications in real world situations, such as the requirement of a large amount of learners and a long period of training cycles in order to discover good learning paths for learners. These practical issues motivate this research. In this paper, the aim is to resolve the aforementioned issues by devising more efficient algorithms that basically run on the same ant colony model yet requiring only a reasonable number of learners and training cycles to find satisfactory good results. The key approaches to resolving the practical issues include revising the global update policy, an adaptive search policy and a segmented-goal training strategy. Based on simulation results, it is shown that these new ingredients added to the original knowledge extraction algorithm result in more efficient ones that can be applied in practical situations.

As the explosion of information due to the Internet in modern age, it becomes more important and more difficult to retrieve information adapted to user preferences. Therefore, personalized recommendation systems are in need to provide proper recommendations based on users’ requirements and preferences (Riecken, 2000). Besides, recommendation systems also find their applications in the eLearning area, especially as personalized web-based learning has become an important learning paradigm in the 21st century. Since learning resources grow so abundant on the Internet that the problem of how to help learners get appropriate learning materials to fit their learning needs has become a popular research subject in the area of adaptive content delivery. A challenging work of this research lies in how to discover effective recommendation knowledge efficiently from past access history of a web-based learning platform (Mor and Minguillon, 2005 and Tang and Mccalla, 2004). An earlier research (Wang & Lin, 2010) showed that a fuzzy knowledge extraction model can be established to extract recommendation knowledge for personalized web-based learning by discovering effective learning paths from a past access database through a novel approach that is inspired by the experience sharing mechanism of natural ants. This approach imitates the natural ants, which share the paths they have found leading to food by scattering pheromone along the paths. In this approach, learners play the role of ants, scattering trail marks in a proper way according to their learning performances along the learning paths characterized by specific learning contexts. These trail marks can then be used to discover effective learning paths for learners with specific learning styles and competency. However, though the research results (Wang & Lin, 2010) revealed the theoretical potential of the proposed method in discovering effective learning paths for learners, critical limitations arose when considering its application in real world situations, such as the requirement of a large amount of learners and a long period of training cycles in order to discover good learning paths for learners. That requirement is impractical for large classes in life and a reasonable time for a course to accumulate good learning experiences should be as short as possible. Therefore, this research is aimed to resolve the aforementioned issues by devising more efficient algorithms that basically run on the same ant colony model yet requiring only a reasonable number of learners and training cycles to find satisfactory good results.

In this paper, an improved ACO algorithm is devised and applied to handle the problem of personalized recommendation for Web-based learning. It was verified through simulations that the segmented-goal training and meta-control strategies proposed in this paper are useful for extracting more effective personalized recommendation knowledge, requiring a smaller amount of learners and training cycles. This makes our algorithm more suitable for practical situations. However, the simulation results also reveal a limitation of the proposed method. In particular, though the quality of the best-found paths can be as high as 0.9, the average effectiveness of virtual learners is still under 0.5. This is because some virtual learners keep exploring the search space in a random way to search for a better path, even when good-quality paths are found. This implies a problem for sharing good experiences: how really good experiences can spread more quickly to stop virtual learners from unnecessary random exploration of the search space. One possible solution to this problem is to cut-off unnecessary parts of the space during the searching process so that the search space can shrink quickly, and the searching is dominated by the best paths. A preliminary result of Wang (2011) has demonstrates the potential of this approach. Another future research is to conduct experiments on real students to investigate its effectiveness in sharing good learning experiences among learners.