针对国内外数字资源服务绩效评价过程中存在的局限性,以及考虑到数字资源服务绩效评价涉及的内外部因素较多,单纯的运用定性或定量评价方法都难以做到准确、客观和全面,本文提出了基于灰色关联分析和BP神经网络算法对数字资源服务绩效进行评价研究。首先建立了基于GRA-BPNN的评价模型,运用灰色关联分析法对评价指标体系进行约简,消除冗余指标,然后把约简后得到的评价指标输入到BP神经网络中进行智能训练,最后把待评价的检测样本输入到训练好的BP网络中,得到了数字资源服务绩效的实际输出值,实际输出与期望输出结果相吻合,从而证明了评价的可行性和有效性。
In view of the exiting limitations in the process of digital resource service performance evaluation at home and abroad,and that evaluation will not be accurate and objective by purely qualitative or quantitative method because of many internal and external factors,The paper decides to evaluate on digital resource service performance based on grey relational analysis and BP neural network.Firstly,the digital resource service performance evaluation model is established based on GRA-BPNN,and reduces its evaluation index system by grey relational analysis,and eliminate redundant indexes,then inputs the reduction indexes to BP neural network for intelligent training.Finally,inputs waiting evaluation test samples to the trained BP network and gets the actual output values of digital resource service performance.The actual output results are matched with desired output results, which proves the feasibility and effectiveness of evaluation.