论文部分内容阅读
首先在结合珠三角民营企业实际情况及劳资关系理论基础上,确定了预警指标体系,并对指标进行筛选。其次,通过主成分分析对指标数据进行降维,得到G1“收入保障因子”、G2“工作环境因子”及G3“劳动合同——争议因子”作为预警模型的输入层。然后,计算劳资关系综合得分值G并对企业劳资关系进行重分类为2(重警)、1(轻警)、0(无警)作为预警模型的输出层。最后,运用神经网络建立劳资关系预警模型,共选取珠三角地区9个城市32家民营企业为样本,以调查问卷的形式获得企业劳资关系满意度数据。结果表明,基于主成分——神经网络(PCA-ANN)的预警模型准确性较高,其中劳资关系重警企业6家,集中分布在佛山地区。
First of all, based on the actual situation of private enterprises in Pearl River Delta and the theory of labor-capital relations, the index system of early-warning was established and the indexes were screened. Secondly, the principal component analysis is used to reduce the dimension of the index data to get the input layer of G1, “Revenue Guarantee Factor”, G2 “Work Environment Factor” and G3 “Labor Contract - Dispute Factor”. Then, calculate the comprehensive score G of labor relations and reclassify the relationship between the labor and capital as 2 (heavy police), 1 (light police), 0 (no police) as the output layer of the early warning model. Finally, the neural network is used to establish the early warning model of labor-capital relations. A total of 32 private-owned enterprises in 9 cities in the Pearl River Delta are selected as samples to obtain the satisfaction data of the labor-capital relations in the form of questionnaires. The results show that the early warning model based on principal component-neural network (PCA-ANN) has high accuracy, of which 6 are heavy-duty industrial relations enterprises, which are concentrated in Foshan.