Abstract
The 21st century is called the century of network relations. Relation-oriented activities and synchronicity in communications of local people have become highly respected, supported and expected, which contributes to enriching social space. Spatial autocorrelation is a concept that treats the inter-relations between adjacent distributed numerical values in a specific time. The statistical model represents an autocorrelation term with parameter "ro" and weight structure matrix "W" in a regression equation.
In conventional inferential statistics, the mutual independency of the observed value is assumed and spatial autocorrelation must not exist in the observed value itself or its residuals. However, the phenomenon that geography treats is ironically filled by spatial autocorrelation, and serves as the basis for the learning of geography. Also in a social phenomenon, people are considered to have mutual influences. Therefore, the scope of this model is wide.
I applied this technique to the local diffusion process of agricultural products, mutual cooperation of users by messaging with a cellular phone, and collaborative exercise lessons of university students. The results of each analysis agreed with the observed phenomena. However, this model has few practical report examples. In order to respond to the needs of the world, the usability of the model needs to be improved and its advantages continuously appealed.
In addition, I examined the facilitation and possible application of autocorrelation analysis to the data of diffusion processes for agricultural products as previously reported.
In this paper, first, the pure autocorrelation model that omits independent variables is compared with normal autocorrelation results. Second, an expansive model that takes the effects of a local diffusion promotion center into the structure matrix W is applied.
JFL Classification: B41, C21, C51, C63, N55, O3, O13, O18, Q15, R12
