Abstract：Cluster analysis of multiple seismic attributes is effective tool for reservoir prediction. The paper presented a fuzzy C-mean algorithm-based cluster analysis of seismic attributes that can analyze multiple seismic attributes and produce useful characteristic pattern, finding optimum cluster center of separate cluster and giving membership grade of each cluster the sample belongs to. The algorithm is simple iterative process, the key of which depends on two parameters: the one is the number of cluster is far less than the total number of cluster samples; the other is the weighting parameters are moderate, too large or too small is of disadvantageous to improvement of clustering effects. The model test and application of practical data showed the algorithm is characterized by both simple principle and robust algorithm, which has good practical application value.
