偏最小二乘回归模型在高校科技人员类型预测中的应用

Application of Partial Least Square Regression Model on the Prediction of Style of the Scientific Personnel in Colleges

目的探讨某省高校科技人员类型的影响因素,建立偏最小二乘回归模型,对高校科技人员类型进行预测、判别,为今后高校科技人员的培养及其合理使用提供更加可靠、科学的理论依据。方法采用自编调查量表——高校科技人员影响因素调查表对该省内高校科技人员进行随机抽样调查,采用Chronbachsα系数和因子分析对量表的信度和效度进行检验,运用SAS9.1对收集的数据进行偏最小二乘回归分析。结果 Chronbachsα系数=0.781,表明调查表具有较好的内在一致性信度,因子分析结果显示量表同时具有较高的结构效度。偏最小二乘回归结果显示,该模型具有较好的拟合优度,并符合专业上的解释,可为人事、科研和教育部门进一步完善高校科技人员培养机制提供理论参考依据。结论偏最小二乘回归模型作为一种新兴的统计分析方法适合应用于高校科技人员类型预测的研究。

Objective To explore the influence factors of the style of the scientific personne1 in colleges,build the partial least square regression（PLS） model,predict and discriminate the style,and supply more reliable and scientific rationale to the training and intelligent use of scientific personne1 in colleges.Methods We took our random sample with inventory done by ourselves—Inventory of Influence Factors of the Colleges Scientific Personnel in a province,then tested the reliability and validity of the inventory by the Chronbach＇s coefficient and factor analysis,done partial least square regression analysis with SAS 9.1.Results Chronbach＇s coefficient equiled 0.781,which indicated the internal consistency reliability of the inventory was good,the results of the factor analysis indicates the structure validity was good,too.The results of the PLS analysis indicatesd the goodness-of-fit of the model was good,which was fit for the major explanation,the model could be used as theory reference for personnel department,research department and education department to consummate the training of scientific personne1 in colleges.Conclusion The partial least square regression model could be used to predict the style of the scientific personne1 in colleges as statistic method appearing recently.