| Evaluating community from the view of ecology is one of the important studies about community in the geography field, This paper, on the base of researches done by many scholars at abroad or home, sets up the frame made of several different indices according the theories and rules on ecological community, including the scale and structure of communities, dwelling conditions, ecological environment, the safeguard and stabilization of communities. Using GIS and RS to quantify every variable, this paper forms the database for evaluating the ecological level of communities. This thesis, selecting the core urban area of Shanghai city as the sample, on one hand, builds a BP neural network model based on the cluster function to judge the level of every community, according to the results, divides all the communities into five types, further analyzes the differences between the communities of the same type; on another hand, by exploring spatial data analysis, this paper discusses the spatial pattern of the ecological indices, and puts forward some advice about the layout of the communities in the view of the spatial relations.This thesis is organized as follows: In chapter 1 "Introduction", the background of the evaluation of communities is provided and conventional theories and methodologies used in evaluating the ecological community are examined, to the existing questions, the study's value is brought out. In chapter 2, methods used in this study and the sketch of the thesis is discussed.Chapter 3 is devoted to selecting variables used to describe the ecological conditions, mainly aiming to be able to quantify every index by making full use of GIS and RS, then the database is created including raster and vector structure. Analyzing the pattern of some individual variable, we can find out some distributing rules for communities' ecological level in given aspect.Chapter 4 is mainly about the evaluating model selecting the core urban area of Shanghai city as an example. Firstly, through factor analysis, five integrated elements are dragged out from all the variables including the scale and structure, the hierarchy of culture, dwelling conditions, ecological factor, green space and the ecological level of the aspects selected is judged. Secondly, BP neural network model is built to compute the ecological index, based on the results, five types of ecological community are divided, and the difference between the communities of the same type is discussed.Chapter 5 is an exploratory spatial data analysis of community's ecological index. It revealed strong evidence of spatial autocorrelation as well as significant patterns oflocal spatial association. The detection of spatial clusters of high versus low community's ecological index distribution throughout the sample period serves an indicator of the persistence of spatial disparities over the ecology landscape. What's more, some advice of optimizing the distribution of the communities is given from the aspect of spatial association.Chapter 6 is about the conclusions drawn from the study; innovations and shortcomings for this study are also put forward.In this thesis, some interesting results are derived from the study, firstly, not only some individual aspect but also the whole ecological index of the communities takes on peculiar spatial pattern, some communities located in the inner area, whose scale of population is very large, are of low level, but the communities staying in the exterior area are at the high level. Secondly, the ecological index of the communities is relating to each other in the space, that is, one area's ecological level is always affected by the neighborhood, most communities take on the similar ecological pattern as the neighborhood, but minority of the ones are different from the others next to them. Thirdly, from the aspect of the spatial association, when come to the distribution pattern of the communities, the existing outline can be ignored, mainly thinking about their ecological similarities in every aspect.
|
PDF Full Text Request