| The fuzzy clustering analysis is a method of discussing quantity of classification kind from the fuzzy set viewpoint. As a multivariate analysis method,it has already been applied in the pattern recognition, the knowledge gain and many other domains. The cluster analysis is a sample classification question under the condition without training regulations. The sample does not know its respective category, but comes automatically on the classification according to the similar degree of sample. According to the similar degree of the pattern characteristic which needs to be classified, the similar is a kind.This paper detailed introduce the basic principle of three kind of fuzzy clustering analysis methods which are the "transmitted the package" based on the fuzzy equal relations, the fuzzy ISODATA cluster based on the soft division and self-organizing iteration. Taking full use of the advantages of this three methods, we propose a kind of method of the classified oil and gas recognition, which combines the K to L transformation with the above two fuzzy clustering methods. The principle based on fuzzy mathematics adapts the fuzziness of logging curve and eliminates the correlation of all kinds of logging parameters. From numerous logging parameters, we can extract classification to the best characteristics of classification.At present, with the development of the technology of seismic attributes, the seismic attributes that we can use increase, this brings some puzzles when we choose the seismic attributes to predict the reservoir. So, if we can make classification of the extracted seismic attributes, then the related will be a class, this will give us certain convenience when choosing attributes.Meanwhile,the data of classifaction is often fuzzy, so in the processing of the data with larger fuzziness, the effect of fuzzy clustering analysis is often better than ordinary clustering analysis. Therefore, the fuzzy clustering analysis is widely used in geophysical exploration areas. |
PDF Full Text Request