Research On Evaluation Model Of Electronic Resources In University Library

In recent years,university libraries continue to strengthen the network and digital development,electronic resources gradually developed into a key part of library document resources.Due to the great development of electronic resources in university libraries,the evaluation and optimization of electronic resources have become particularly important.However,at present,the evaluation of electronic resources is still based on subjective decision-making,which lacks objectivity and scientificity,and is not conducive to the construction of library resources.Based on this,this paper proposes an evaluation index system and model of electronic resources,in order to improve the scientificity of evaluation work and promote the optimization and value playing of electronic resources.The specific work is as follows:Firstly,sum up other scholars in the field of electronic resources evaluation results,combined with the actual purchase and usage of electronic resources in university library research,and the characteristics of the discipline construction,analyses the influence factors of the electronic resources evaluation.What's more,gray statistical theory is used to screen key indicators,and a new electronic resource value evaluation system is constructed,which includes 10 indexes,such as resource type,resource quantity,core periodical inclusion rate,retrieval function,probation period,key subjects,user satisfaction,annual price increase,per capita usage,and usage cost.Secondly,K-means algorithm is used to generate category tags based on the internal similarity of electronic resource data,divide electronic resources intodifferent categories,and evaluate and analyze each category respectively,so as to transform the subjective evaluation of electronic resources into an objective classification problem reasonably.Thirdly,according to the idea of gradual optimization,the evaluation model of electronic resources was established by using CART algorithms and RF algorithms respectively,and the RF algorithm was further optimized by using grid search to establish the evaluation model of electronic resources based on GS-RF algorithm.Finally,through comparison experiments to verify the effects of the above three models,it is found that the GS-RF algorithm based university library electronic resource evaluation model has the best prediction effect,with an accuracy rate of94.44%.The research results of this paper can provide some theoretical guidance and decision support for the evaluation and optimization of electronic resources in university libraries,and provide a new way of thinking and methods for subsequent research.