Evaluating industrial cluster by using spatial auto correlation of patent applications

Development of an industrial cluster that denotes a geographic concentration of interconnected businesses and associated institutions in a particular field is one of most important policies for many countries. One of the key issues for promotion of the policy is to use its proper assessment. At the present time, some assessment methods based on economic statistics or questionnaire investigations are proposed. However, the methods based on economic statistics can apply to only longterm assessment. On the other hand, the methods using questionnaire investigation include problems of consuming a lot of time and effort. In order to solve the problems, we develop a patent analysis method which uses geometric bias of patent applications, which is able to apply for middle/short-term assessment by using global Moran's test and local Moran's test that measures spatial auto-correlation. As a result, our method can detect the bias on patent applications.