权重矩阵设计对价格空间属性判断的影响

发布时间：2018-05-19 02:36

本文选题：空间权重矩阵 + 空间自相关性　；参考：《青岛大学》2017年硕士论文

【摘要】：空间权重矩阵的选取问题是空间统计学中一个重要的研究热点。研究发现,空间权重矩阵不同,商品价格空间关系结果不尽相同,甚至差异较大。对商品价格空间关系认识不清,不利于商品价格管理、市场整合等。如何正确选取并设计空间权重矩阵,进而对商品价格的空间属性进行较为科学合理的判断,是目前研究领域的薄弱之处。本文首先对历年国内外空间权重矩阵研究文献进行梳理。其次,对多类空间权重矩阵的理论知识进行阐述,对文中可能涉及到的其他理论知识进行概述。然后,利用不同种类的空间权重矩阵对食品类、衣着类、住房服务类、交通服务类四大类中的具体商品进行实证分析,重点在于比较不同空间权重矩阵对各类产品的影响及差异,达到对产品价格空间属性进行判断的目的。最后,利用空间计量模型对商品的影响因素进行分析,进而得到其传导机制,得到的基本结论如下:(1)对不同权重矩阵进行参数优化时,每一种权重矩阵都有不同的构造原理与影响因素,对不同种类产品空间关系的影响不尽相同。概况来说,适用范围从大到小依次排序为:复合权重矩阵、距离权重矩阵、邻接权重矩阵、高阶权重矩阵。不同的商品,判断空间属性的最优权重矩阵也不尽相同。(2)对于食品类、衣着类、住房服务类、交通服务类这四个类别的商品来说,食品类中大多数的蔬菜类商品与养殖类商品具有空间自相关性,而该类别其他商品空间自相关性较弱。衣着类、住房服务类和交通服务类商品几乎不存在空间自相关性。(3)如果将商品分为政府限价商品及非限价商品,可以发现政府非限价商品大部分具有较强的空间自相关性,少部分相关性较弱,原因这些产品无论是通过网络途径还是实体店途径,各地价格差异相对较小。政府限价商品的空间自相关性相对于政府非限价商品来说要低,且政府限价商品中可流通商品的空间自相关性要强于不可流通商品。(4)结合实际情况,借助空间计量模型发现具有空间属性的商品价格主要是受交通成本以及替代品价格的影响程度较大。
[Abstract]:The selection of spatial weight matrix is an important research hotspot in spatial statistics. It is found that the spatial relationship of commodity prices is different or even quite different with the spatial weight matrix. Not clear understanding of the spatial relationship of commodity prices, not conducive to commodity price management, market integration and so on. How to correctly select and design the spatial weight matrix and then make a more scientific and reasonable judgment on the spatial attribute of commodity price is the weakness of the current research field. In this paper, the domestic and foreign literature on spatial weight matrix is firstly reviewed. Secondly, the theoretical knowledge of multi-class spatial weight matrix is expounded, and the other theoretical knowledge that may be involved in this paper is summarized. Then, using different kinds of spatial weight matrix to analyze the specific commodities in the four categories of food, clothing, housing services and transportation services. The emphasis is to compare the influence and difference of different spatial weight matrix on all kinds of products in order to judge the spatial attribute of product price. Finally, the influence factors of commodities are analyzed by using spatial econometric model, and the transmission mechanism is obtained. The basic conclusions are as follows: 1) when the parameters of different weight matrices are optimized, Each weight matrix has its own structure principle and influence factors, which has different influence on the spatial relationship of different kinds of products. Generally speaking, the order of application range from large to small is: compound weight matrix, distance weight matrix, adjacent weight matrix, high order weight matrix. For different commodities, the optimal weight matrix for spatial attributes is not the same.) for food, clothing, housing services, transportation services, these four categories of goods, Most vegetable commodities in food category have spatial autocorrelation with aquaculture commodities, while other commodities in this category have weak spatial autocorrelation. Clothing, housing services and transport services have almost no spatial autocorrelation.) if the goods are divided into government limited price goods and non-price goods, it can be found that most of the government non-price goods have strong spatial autocorrelation. A small part of the correlation is weak, the reason for these products whether through the network or physical stores, the price difference is relatively small. The spatial autocorrelation of government limited price commodities is lower than that of government non-price limited commodities, and the spatial autocorrelation of government limited price commodities is stronger than that of non-tradable commodities With the help of spatial measurement model, it is found that the commodity price with spatial attribute is mainly affected by the transportation cost and the price of substitute.
【学位授予单位】：青岛大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F014.31

【参考文献】