Research On Ontology - Based Learning Resource Construction Model And Its Application

In the information age of the 21st century, internet provides convenience for our learning. Based on the information environment, the way of learning is more inclined to multimedia learning, E-learning and Mobile Learning, these ways are inseparable from a huge information repository which the Internet provides. However, information resources on the network covering a wide range, scattered distribution, large capacity, Coupled, Network make everyone can obtain and store information on the internet, and no quality control and management mechanisms, so that, network filled with a lot of useless information. When users find information, so many resources are retrieved, and present in different forms. Therefore, in the information explosion era, user want to find resources to meet their needs with a certain challenge. Therefor, how to build resource repository techniques which used to meet user themselves query needs and resources recommendation、query algorithm based on user personalized demand has become a hot research.General resource repositories can build manually or automatically. Manually building can use field’s key words to store resource information in resource repository, building a comprehensive resource repository. But with the expansion of the field, keywords will be more and more, corresponding resources are more abundant, manually building model becomes time-consuming and laborious, can’t meet the requirements. For time-consuming problem, automatic building has great significance. Automatic building model implements based on the web crawler algorithm, learning resources in resource repository are independent without any contact. For personalized resource push and retrieval will ignore the semantic issue, that is likely to ignore the user’s real needs and the true intentions of the queries, this will make resource recommendation and query not accurate in result. In this paper, using the PageRank algorithm to fetch web resources while combined with domain ontology to analyze semantic correlation among resources, automatically build learning resource repository, and apply to personalized resources recommendation.The main contents of this paper are as follows:(1) Automatically build and store learning resource repository based on ontology. For existing resource repository’s shortage, this paper propose a learning resource building model based on domain ontology and web crawler algorithm, this model using PageRank algorithm to grab web resources, combined with domain ontology to analyze semantic correlation among resources, and build resource repository in some field. While analyzing and comparing existing ontology storage methods, using MySQL relational database to store according to some certain rules.(2) Resource of user interest model. Due to information displays in different ways and is has different effects on user interests, improving the general user interest model according to the resource type, mainly with interested content and resource type. Analyzing and improving the weight estimation of user interest and analyzing resource type in user personalized needs, and using forgetting factor to update the concept of user interest.(3) Personalized resource recommendation. Combined with domain ontology and Ant Colony Algorithm, this paper propose a personalized recommendation algorithm based on domain ontology and Ant Colony Algorithm. Firstly, using Ant Colony Algorithm cluster users, getting several great similarity users; Secondly, extracting its interested concepts according to the similar users, composed of a set of concepts, recommending resources from learning resource repository, completing personalized resource recommendation, and verified by experimental system.