Font Size: a A A

Evaluation System BP Network Of Networks Based On Shared Scientific Literature

Posted on:2015-02-03Degree:MasterType:Thesis
Country:ChinaCandidate:S J FengFull Text:PDF
GTID:2268330428977778Subject:Systems Engineering
Abstract/Summary:PDF Full Text Request
Science and technology literature resource is the important guarantee ofnational science and technology innovation development, and the constantlyupdate and popularization of Internet technology in modern times, has formed ahuge demand for network science and technology literature resource. In order tomeet users’ needs, improve the service of network science and technologyliterature platform, promote the construction and management of networkscience and technology literature platform, and make the effective use ofresources rationally, it needs to make a scientific and objective evaluation of thesharing degree of network science and technology literature resource from theusers’ perspective.First analysis of the current evaluation methods commonly used indicatorsystem and the characteristics evaluated, select the BP neural network algorithmas a research tool in this paper. Secondly, according to the characteristics andconnotations of the network science and technology literature resource sharing,puts forward the evaluation model of "Visibility, Availability, Usability". Aroundthe evaluation model, it analyzes the users’ experiences and feelings aboutachieving science and technology literature resource and practicing through thesharing platform, through the method of literature review, group discussion andexpert consulting, it establishes the structure of the primary indexes forevaluation of network science and technology literature resource sharing. Usingthe PSO algorithm to optimize the BP neural network algorithm when training inlimitations exist, formation of PSO-BP neural network model; Collect thenetwork training sample data by questionnaire survey, and build the constructionof PSO-BP neural network training model. Input the sample data, and train theestablished model. According to the link weights between each layer neurons,the index weight of science and technology literature resource sharingevaluation index system is obtained, and it analyses the weight of each index inthe index system, perfects and establishes evaluation index system of network science and technology literature resource sharing.Finally, in the case of theevaluations of provincial and municipal local network science and technologyliterature resource sharing platform, by making an instantiated applicationresearch, and doing a test evaluation and analysis according to the establishedevaluation index system, which verified the validity of this paper work.
Keywords/Search Tags:Network S&T literature, Resources sharing, Evaluation indexsystem, BP neural network
PDF Full Text Request
Related items