基于面板数据的接近性和相似性关联度模型

对于面板数据, 首先给出面板数据的空间投射方法, 将面板数据投射为空间的向量序列. 然后, 基于空间向量的夹角和距离分别构建相似性和接近性关联度模型. 具体方法为: 利用向量夹角构建面板数据的相似性关联度模型; 同时, 基于向量差的模构建面板数据的接近性关联度模型. 分别讨论两种关联度模型的规范性和对称性等性质. 最后, 通过实例验证了相似性和接近性关联度模型的合理性. 分析结果表明, 所提出的关联度模型能较好地反映面板数据的相似性和接近性的关联程度.

Similarity and nearness relational degree based on panel data

This paper firstly presents the projection method of the panel data, which is projected to space vector series. Based on the angle and distance of the space vector , this paper respectively constructs the similarity and nearness relational degree models based on the panel data. The angle of the vector is used to construct the similarity relational degree model based on the panel data. Meanwhile, the distance of the vector is used to set up the nearness relational degree model based on the panel data. The properties of two models, normalization and symmetry, are also discussed. Finally, an example is given to illustrate the rationality of the similarity and nearness relational models, which suggests that the proposed models can reflect the similarity and nearness relational degree of the panel data.