The Application Research Based On Fuzzy Clustering And Trend Surface Analysis

With the development of China’s modernization, more and more geological exploration projects are underway, a large number of geological survey data need to be processed analysis, trend surface analysis algorithm is a mathematical calculation of geological analysis, which calculated a function species-specific study to analyze the spatial distribution characteristics, as well as the impact and trends. However, in practical problems, there are certain limitations computing trend surface in exceptional circumstances as the traditional method of least squares method to solve, such as when the control points are distributed unevenly, increasing the number of times when the trend surface, easy to appear normal equations sick, for the mathematical concept of fuzzy clustering analysis among them the introduction of trend surface analysis, will greatly enhance the results of the stability and convergence.ISODATA cluster analysis algorithm is based on the k- means algorithm based on the increase of the clustering results for "consolidation" and "split" the two operations, and set a clustering algorithm run control parameters. ISODATA fuzzy clustering analysis to the equation is solved when the fuzzy set different trends surface coefficient constituted sample its corresponding measured values are formed among this collection, we have the entire collection is divided into two different subsets, part is relatively close to the standard set of samples, in part, relatively far in the standard sample collection, in the calculation process, each calculation have increased closer to the true value of a sample, remove the sample does not fit, can be calculated by repeatedly make a collection of cluster centers closer to the true value.The issue in the main line trend surface analysis algorithms, fuzzy theory points, highlighting the principle and application of the results. The main contents of the following aspects:1) The trend surface analysis algorithm with fuzzy clustering algorithm is the basic principle of the least squares method computing trend surface analysis and ISODATA clustering algorithm to calculate the coefficients of trend surface analysis process described, give a specific solution procedure and trend surface in tended test combined degree.2) Differences in the basic algorithm of trend surface analysis and algorithm analyzes using fuzzy elective principle evaluate trends at various points near surface extent of impact on the overall trend.3) Application of surface analysis of trends algorithm which uses specific examples, given differences in a more detailed comparison of different methods of analysis and the application of results-based ISODATA clusters.