Research Of Factor Analysis For Overall Well-off Society Based On Improved GMDH Method

In this research, we concentrate on developing a Group Method of Data Handling model for a county. We first assign a set of index system to evaluate the well-off degree, then for each index we collect some influence factors that may affect this index. By modeling with GMDH method, we can get equation for each index, which reflects factors that influence each index. These equations can provide scientific evidence to well-off development in county area, therefore we can offer some policy suggestions for the county.An overall well-off society is China's grand blue print in the first two decades this century, also a complex social project. Based on the well-off construction in a county of Sichuan Province, by the means of system model group and self-organization data mining method, we deep into it's well-off process, motivating factors and hampering factors, and provide scientific planning and strategy choice basis.The well-off evaluating system involves many measuring factors, and each of them is directly or indirectly related with economic, social, environmental area. We need to use the special method which is appropriate for researching complex system like this. So, we choose GMDH method.The group method of data handling was introduced in the early 1970s. GMDH is a set of several algorithms for different problems solution. It consists of parametric, clusterization, analogs complexing, rebinarization and probability algorithms. This self-organizing approach is based on the sorting-out of gradually complicated models and selection of the best solution by minimum of external criterion characteristic. Not only polynomials but also non-linear, probabilistic functions or clusterizations are used as basic models.When we have got the analysis models for every factor, we take use of fuzzy multi-objective optimization method to testify whether the models we got by GMDH method is accurate. And the optimization results we got from fuzzy multi-objective optimization method is similar to the real data, so we can believe the results we got from GMDH method reflect the real relationship between the factors and their influence index.