Spatial dependency describes the relationship between one dependent spatial variable and other related spatial
variables. This paper constructs two kinds of Fuzzy Neural Networks for spatial dependency mining, the modified fuzzy
neural network model and the fuzzy comprehensive assessment network model. The first model is built from general
fuzzy neural network model. It has four layers, input layer, fuzzy membership function layer, fuzzy reasoning layer and
output layer. The second model is built based on a fuzzy comprehensive assessment algorithm. It has five layers. The
first three layers are same as the first model, the fourth and the fifth layer are used to find the maximum membership
degree and give the output. We develop the training algorithm for these two models based BP algorithm and genetic
algorithm, respectively. This paper adopts a thematic spatial database of land evaluation to test these models. We use
experiential knowledge as original rules to build initial FNN models. We can see that original rules (spatial dependencies)
are corrected after training. It can be seen that these two models get almost the same revised dependencies, and this
indicates that these two models both correct the original ones and get the more objective spatial dependencies.
Experiments also indicate these two models are efficient.
This paper studies the principle, method and application of spatial points clustering based on self-organizing neural
networks. In this paper, we put forward a kind of composite clustering statistic, called generalized Euclidean distance,
which is calculated by both geometric and semantic characters of spatial points. We propose the algorithm of spatial
points clustering based on self-organizing feature map and generalized Euclidean distance. The clustering method in this
paper can generate better result reflecting the clustering characters of spatial points. Finally, we employ a case study to
probe into data classifying, gross error detecting and homogeneous areas partitioning using self-organizing spatial
This paper focuses on application of artificial neural networks (ANN) in land suitability evaluation. There are some problems in applying fuzzy system to land suitability evaluation such as self-adjusting ability of the membership functions and rules of fuzzy evaluation system. In this paper, the model of fuzzy neural network is designed for land suitability evaluation. This model is the result of integrated fuzzy system and artificial neural network. This fuzzy neural network model has five layers. The learning algorithm of the model has been designed based on the principle of error back propagation of neural networks. The learning strategy, algorithm and efficiency of the model have been tested and the results of test are satisfied.
According to the specific situations in China, this paper discusses the application of RS data in analysing the dynamic balance between cultivated land supply and demand which is one of main tasks of the Ministry of Land and Resources of China. It points out that based on applying RS data to monitoring land use changes, we can make full use of RS data to extract the information required in the analysis on the balance, which is an important approach for dynamically mastering and regulating the balance. It presents the framework and main aspects for analysing the balance, including the environment of the balance, the elements of the balance, the state of the balance and the process of the balance, as well as analysis on the balance at multimeasures, such as the balance in quality, in Gross Amount, in Per capita Amount, in Region and in Time.
Artifical Neural Networks (ANN) has many good qualities comparing with ordinary methods in Land Suitability Evaluation. Based on analysis of ordinary methods' limitations,s ome sticking points of BP model of ANN used in land evaluation are discussed in detail, such as network structure, learning algorithm, etc. The land evaluation of Qionghai city is used as a case study, we know that ANN always can give more reasonable evaluation results from test.