基于GA-BPNN组合模型的图书馆绩效评价及实证研究  

Library Performance Evaluation and an Empirical Study Based on GA-BPNN Combinatorial Model

作　　者：张秀华[1] 高伟[2] 

机构地区：[1]鲁东大学图书馆,烟台264025 [2]鲁东大学数学与信息学院,烟台264025

出　　处：《情报学报》2012年第8期890-896,共7页Journal of the China Society for Scientific and Technical Information

基　　金：2011年教育部人文社会科学研究青年基金项目（编号：11YJCZH039）;山东省2010年人文社科项目（编号：J10WL51）.

摘　　要：针对国内外图书馆绩效评价过程中评价方法存在的局限性，以及绩效评价系统指标因素的模糊性、不确定性、难以量化等特征，提出了遗传算法和BP神经网络算法相结合的GA—BPNN组合模型对其进行评价，首先建立基于GA—BPNN的评价模型，运用遗传算法GA优化BP神经网络的连接权值和阈值，然后把优化好连接权值和阈值输入BP网络进行智能网络训练，最后把待评价的测试样本输入到训练好的BP网络中进行实证分析，得到图书馆绩效评价的实际输出值，与传统BP神经网络算法得出的结果对比，拟合精度、准确度、效率大幅提高，从而证明该模型具有较好的可行性和实用性。In view of the exiting limitations of evaluation method in library performance evaluation at home and abroad, and its index features are difficult to measure by exact digital indexes. The paper decides to establish GA-BPNN combinatorial model to evaluate it. Firstly, GA-BPNN evaluation model is established, and optimizes the connect weights and thresholds of BP neural network by genetic algorithm, then inputs the optimized weights and thresholds to BP neural network for intelligent training. Finally, inputs the waiting evaluation test samples to the trained BP neural network for empirical analyzing and gets the actual output values of library performance. The method gets higher precision, accuracy and efficiency of simulation by comparing with national BP neural network, which proves the good feasibility and practicability of the model .

关 键 词：遗传算法 BP神经网络 图书馆 绩效评价 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]