Intelligent Tutoring System Based On Learner Characteristics Extracted

Intelligent Tutoring System intelligent simulated the behavior of teachers to guide learners on the network one to one targeted learning. Learner model is the core of Intelligent Tutoring System and the key of the system's practicability is how to acquire behavioral character of individual learners. Using association rule mining algorithm can effectively extract the behavioral character of learners, thereby enhancing the Intelligent Tutoring System automatically recommend targeted.This paper constructs the learners'feature model for Intelligent Tutoring System based on analyzing traditional network learner model, we use association rule mining algorithm to extract the behavioral character of learners in this model. By the research of classical association rules Apriori algorithm, we found this algorithm scan the database every time when the frequent sets are generated, that reduce the efficiency in large transaction database. To solve this problem, we propose the Flag_Apriori algorithm correspondingly, that separates the transaction records having no relevant to the generation of frequent sets by setting flag and the scanning for those records will be ignored. Examples on SQLServer show this algorithm is efficient than the Apriori algorithm, so that it is faster when we mine the character of learners in large-scale data.Base on the research above, we design an Intelligent Tutoring System prototype in LAN. According to the learner's personal learning character, it automatically recommend relevant pages that learner may be interested in to provide personalized recommendation service in this system.