The Effect Of Weight Matrix Design On The Judgment Of Price Space Attributes

The selection of spatial weight matrix is an important research topic in spatial statistics.The empirical results indicate that the commodity price spatial relationship results are not the same,even the results are quite different,because of the different use of the spatial weight matrix.The unclear understanding of the commodity price spatial relationship is not conducive to the management of commodity prices and market integration.How to select and design the spatial weight matrix,and then makes a scientific and reasonable judgment on the spatial attributes of commodity prices,is the weak point in the current research field.Firstly,this paper combs the research literature on spatial weight matrix.Secondly,this paper describes the theoretical knowledge of varied spatial weight matrix,and other theoretical knowledge that may be involved in the paper.Then,in order to achieve the purpose of judging the product price space attributes,empirical analysis is given on the basic necessities of life related products in four categories by using different kinds of spatial weight matrix.The key point is to compare the influence and difference of different spatial weight matrix on various products.Finally,using the spatial econometric model to analyze the influencing factors gets the transmission mechanism.The basic conclusions are as follow:(1)When the parameters of different weight matrix are optimized,each weight matrix has different construction principles and influencing factors,the impact of different types of products on the spatial relationship is not the same.In general,the scope of application followed in descending order: composite weight matrix,distance weight matrix,adjacency weight matrix,high order weight matrix.And the optimal weight matrix is different for different commodities.(2)For the above four categories,most of the vegetables and aquaculture products in food commodities have spatial correlation,but the other category of commodity space correlation is weak.There is little spatial correlation among clothing,housing and transportation services.(3)If the goods are divided into government limit commodity and non-limit commodity,most of the government non-limit price commodities have strong spatial correlation,but the others have weak correlation.The reason is whether the product through the network or the store sales,price differences are relatively small.The spatial correlation of government price fixing commodity is lower than that of government nonlimit commodity.Moreover,the spatial correlation of the negotiable commodity is stronger than that of the non negotiable goods.(4)Using spatial econometric model,it is found that the price of goods with spatial attributes is mainly affected by the cost of transportation and the price of substitutes.