基于投影寻踪模型的图书馆成效(绩效)评估研究

[目的/意义] 基于已有的评价指标体系，建立图书馆成效（绩效）评估的投影寻踪分类（PPC）模型，厘清各个评价指标与图书馆成效（绩效）之间的影响关系及其重要性，对图书馆成效（绩效）高低进行排序，提出提升图书馆成效（绩效）的措施和建议。[方法/过程] 根据抽样得到的图书馆实际调查评价指标数据，编制基于群智能乌鸦搜索算法的PPC模型程序，选取合理的窗宽半径值，建立求得真正全局最优解的图书馆成效（绩效）评价PPC模型、投影向量及其系数和样本投影值。[结果/结论] PPC模型能够同时完成确定各个评价指标的权重和构建图书馆成效（绩效）评价函数，具有客观性好、结构简单、数学意义清晰和后续应用便捷等特点，评价结果与专家评价法结果基本一致，说明PPC模型能很好地应用于图书馆成效（绩效）评估研究，为图书馆成效（绩效）评估提供新方法。

The Library Outcomes (Performances) Evaluation Using the Projection Pursuit Clustering Model

[Purpose/significance] Based on the established evaluation indexes for the library outcomes (performances), the projection pursuit clustering (PPC) model is applied to evaluate the library outcomes (performances), to clarify the influence relationship between the evaluation index and the library outcomes (performances), to rank the library outcomes (performances), and thus to put forward the measurement improving the library outcomes (performances).[Method/process] According to the collected sample data of the library outcomes (performances), the Matlab program of the PPC model based on swarm intelligence crow search algorithm is developed. The reasonable cutoff value is taken. The PPC model of the library outcomes (performances) with the real global optimization, the projection vector as well as its coefficient, and the projection value of the samples are established.[Result/conclusion] The projection vector coefficient of each evaluation index and the evaluation function or the scores of the library outcomes (performances) are simultaneously obtained by optimizing the PPC model. The PPC model possesses the characteristics of the simple structure, the good objective, the clear mathematical meaning, and the good convenience to apply the PPC model. The rank and the classification of the library outcomes (performances) using PPC model are good agreement with the results according to the experts. It is approved that the PPC model is applicable, effective and a new method to the library outcomes (performances) evaluation.