作者：李顺勇；岳利梅

第一作者单位：山西大学数学科学学院

摘要：为研究结构方程模型的偏最小二乘回归在成分数据中的应用,利用旅游业投资、接待能力以及发展三者之间的关系进行模型分析。通过将偏最小二乘回归应用于成分数据中,提出了新的修正加权符号权重和,进而得出潜变量的估计值。研究结果表明,通过修正后的加权符号权重和,使得模型中显变量的拟合效果更好,且通过分析相关系数可知新模型结果与实际情况更符,即旅游投资主要通过影响旅游接待能力促进旅游业发展。

Abstract：In order to study the application of partial least squares regression of structural equation model in the compositional data, this model is analyzed through the relationship among investment, reception capacity and development of tourism. Based on the analysis of available PLS in compositional data, a new weight of sign weighted sum is proposed, and the estimated value of the latent variables also can be obtained. The results show that by correct of the weighted sign weighted sum, the fitting effect of the explicit variables in the model is better, then by analyzing the correlation coefficient, the results of the model are fit with the actual result, that is, tourism investment is mainly through the influence of tourism reception capacity to promote the development of tourism.

基金：国家自然科学基金资助项目(61573229)；山西省自然科学基金资助项目(2015011044)；

关键词：成分数据;结构方程模型;偏最小二乘;符号权重和;旅游接待能力;

DOI：10.16186/j.cnki.1673-9787.2017.03.024

分类号：O212