Building Association Link Network For Managing Large-Scale Web Resources

As a sharing intermedium for global information, Internet has become a powerfor economics and society. In the internet environment with huge number ofresources, different resources cluster as a domain. Thus, the internet has become acomplex information sharing and storage intermedium, which is with domain featureand large scale resources. The resources in different domain are increasing withexponent, which is heterogeneity, weak association, and dynamic. The gap betweenthe users and resources is broader than ever. How to manage and semantic linkresources effectively has become a hot research topic in web information processingfield.In this paper, we propose a web resources organization model based onassociation semantics, i.e., Association Link Network, which focuses on three keyproblems:(1) The semantic discrimination capability on keywords of resources.How to compute the semantic discrimination capability on keywords, which makeskeywords be accord with the cognitive ability of users. How to narrow the gapbetween the keywords and users’ cognitive ability, which makes users search andbrowse web resources effectively and efficiently. Computing semanticdiscrimination capability of keywords is basic for building semantic link betweenresources.(2) The semantic link between resources. How to build semantic linkbetween resources, which transforms the resources from cyber space to semanticspace. How to keep the autonomous, self-organized of resources organization model,which manages the web resources effectively and efficiently.(3) The semantic constraint between resources. How to keep the dynamicsand temporal feature of resources organization model, which makes the resourcessharing accurately and conveniently. How to keep the standard of resourcesorganization model, which meets the cognitive need of web users.In order to solve the above three problems, the major contributions of this paperare as follows:(1) In order to solve the first problem, the web resources are divided into topiclayer, resources layer, and keywords layer. The domain with large scale resourcescan be clustered accurately after dividing into different layers. In the keywords layer,the semantic discrimination capability of keywords is proposed, which is relevant tothe cognitive ability of users. The computation of semantic discrimination canprovide a solid base for building semantic link between resources. Meanwhile, weuse the power law function of keywords to computer semantic discriminationcapability of keywords. The experimental results prove that the proposed method canleverage the accuracy and complexity.(2) In order to solve the second problem, we propose a web resources organization model based on association semantics-Association Link Network.Through building the semantic link between resources, the large scale resources in adomain can be organized effectively. The proposed model builds semantic linkautomatically, do not need ontology, lessen the cognitive burden of users.(3) In order to solve the third problem, we propose incremental buildingassociation link network algorithm. Through collaborative filtering and duplicatedetection technology, we add the temporal feature to association link network, whichmeet the dynamic of web resources. Meanwhile, we present the different schema forbuilding association link network based on complex network, which keeps theautonomous and self-organized of association link network. Moreover, we develop aweb events analyze system which contains6million we pages. Through theevaluation by real large scale domain, the proposed model can build intermedium onkeywords, resources, and topics layer, and provide effective knowledge services forusers.