Individual differences in learning has long been of interest to educators. The idea that individual differences in learning it solely due to differences in intelligence and ability to live long in the world of education was accepted, but later was changed. The researchers showed that students have different learning styles, different means of information and analysis they learn. Today, it's clear that different people learn only partly depend on their intelligence and abilities, so other factors such as personality traits, task difficulty and different styles are also involved. In general, the term light refers to the dominant paradigm person doing things. Learning styles, preferences or usual method for learning. Smart learning process model in the educational system, called a learner model. Learning model includes features such as knowledge, interests, goals, background, emotional behavior, personal characteristics, cognitive characteristics and learning styles. Using these features learner behavior and how to process user information to predict and therefore training services will be provided in accordance with his needs. The study focuses on the characteristics of the learning style. Learning style preferences and individual approach to learning notes. Studies show that taking into account the learning style of education, increase the quality of learning. Conventional methods to extract learning style, using a questionnaire. Inventory problems with it, such as boring, are not changes to individual learning styles, choose incorrect answers mentioned. Automatic detection of the method of learning styles as an alternative to conventional methods used. The feature can be used to extract network of clubs such as feature extraction method, or the characterization of the individual in a manner Duff, or use the nearest neighbor's fuzzy. At the end of a combination of tree structure and fuzzy logic is used to determine the learning style. The proposed method is based on 4596 data from the knowledge base of students that Acer Australia has been extracted, were studied. Using features to reduce PCA, unnecessary features have been removed and then a variety of fuzzy decision tree algorithms have been implemented on these features. The results indicate good accuracy of the proposed method compared to other methods that were used in the decision tree is.
