基于改进AHP-BP神经网络的科研项目数据库评价指标模型构建  

Construction of Evaluation Index Model of Scientific Research Project Database Based on Improved AHP-BP Neural Network

作　　者：黄仕靖 陈国华[2] 吴川徽 袁勤俭[1] HUANG Shi-jing;CHEN Guo-hua;WU Chuan-hui;YUAN Qin-jian(School of Information Management,Nanjing University,Nanjing 210023,China;School of Engineering Management,Nanjing University,Nanjing 210029,China)

机构地区：[1]南京大学信息管理学院,江苏南京210023 [2]南京大学工程管理学院,江苏南京210029

出　　处：《情报科学》2020年第1期140-146,共7页Information Science

基　　金：江苏省社会科学基金项目“学术虚拟社区知识交流的效果评价研究”(17TQB003)

摘　　要：【目的/意义】应用改进的AHP-BP方法构建基于用户感知的科研项目数据库服务质量评价体系。【方法/过程】根据传统AHP法得出指标权重,计算出专家群组的净感知相关系数矩阵,确定了专家权重,进而求解得到评价指标的综合权重,然后以多组数据为先验样本进行BP神经网络的训练、测试和验证,从而得出了可供推广的AHP-BP神经网络的科研项目数据库综合指标权重模型。【结果/结论】对指标权重的分析表明,对于科研项目数据库来说,内容是最关键的评价指标,且改进的AHP-BP神经网络评价模型所得结果更加客观合理。[Purpose/significance]Applying improved AHP-BP method to constmct user service database quality evaluation system based on user perception.[Method/process]According to the traditional AHP method,the weight of the index is obtained,and the net perceptual correlation coefficient matrix of the expert group is calculated.The expert weight is deter・mined,and then the comprehensive weight of the evaluation index is obtained.Then,the BP neural network is trained,tested and verified with multiple sets of data as lhe prior samples,which led to the comprehensive index weight model of the re・search project database of the AHP-BP neural network available for promotion.[Result/conclusion]Analysis of indicator weights shows that content is the most critical evaluation indicator for research project databases,and the improved AHP-BP neural network evaluation model is more objective and reas on able.

关 键 词：科研项目数据库 用户感知 改进AHP-BP 评价指标模型 

分 类 号：G250.2[文化科学—图书馆学]