Fig.6.Automatic learner model [16]
Learner models can be extracted based on Felder– Silverman learning style model personality which consider the learning styles, behavioral factors like user’s browsing history and knowledge factors like user’s prior knowledge. The learner’s are classified based on their interests by using NBTree classification algorithm in conjunction with binary relevance classifier. The keyword mapping part and profile table structure can be enhanced
depending on the application area for a particular learner model. 4.3 Vector space models for user modeling The user model [14] is one in which one collect a few specific facts about the person and then invoke knowledge about the groups to which the person belong. First user modeling system is the recommendation system. Another approach of user modeling is to use vector space models shown in fig.7, where an item is represented in an n- dimensional vector space. A student can be represented in an n- dimensional vector space as feature vector, where each feature corresponds to n individual concept associated with a given course.
A student’s knowledge of the materials in the course can be represented as the feature vector, (knowledge_topic0; knowledge _ topic1; . . . ; knowledge _ topic n). The similarity between two users can be calculated by using the Euclidean distance between two points. K-means clustering or any other technique is used to maintain or update continuously based on new observations about the users.
For updating and maintaining the information in the stereotypes, the triggers attached to each stereotypes position within the hierarchy itself. In vector space models, the representation of individual user is adjusted when new information is reported or observed. Vector space models are easy to store in database tables. The database can be easily accessed, queried and manipulated using structured query language statements and queries.
Fig. 7.User model attributes for each course item[14] This user model has been implemented for the Moodle Elearning environment. Course items viewed, quiz and test attempts and scores, assignment submissions and results,
updates and alterations to course items, times of log ins and log outs are used for every user interaction.
In each class of students a number of distinct student groups were identified to analyze the data retrieved by the user modeling component. The grouping system was developed by modeling each user as a vector model in an n- dimensional space and using a weighted n- dimensional Euclidean distance algorithm to calculate the distance between any two users. Vector space modeling is used to develop the student models in an E-learning system which enhances the effectiveness or usability of the software systems. New principles can be applied to the intelligent user interfaces, so that they provide repeatable and transparent interactions of the users with the system.
5. Conclusions
Now a day’s offering E-learning courses are increasing at an unrestrainedly pace. However, the learning experience is often perceived by the user as a one-way highly constrained communication process, where the computer is only the mechanical device that conveys the content. Self regulated learning [19] help students to develop learning habits. In order to increase student’s learning motivation and to develop practical skills, problem based learning is considered to be one of the most appropriate solution. The art of designing good E-learning systems is difficult and is of great challenge for the human mind. The way this is done is also dependent on the learning culture in each country. The key issue is to facilitate new learning modalities for younger generations. Future Investigating methods in an E-learning system are to support students with special needs such as super intelligent, retarded, etc. To develop that knowledge-.bases (ontologies and LORs) which are available automatically from instructor’s submitted multimedia learning material
References