The Improvement Of K-Means Algorithm And Its Applications In Analysis Of Geological Exploration Seismic Data

Clustering is a method of data partition or grouping, the method is an important area of data mining. K-means algorithm is an algorithm that is based on partitioning method and is one major branch of the clustering algorithms.The Oil is a national strategic materials, and most important with people’s lives. The domestic oil production is in the age of post-mining, so the oil companies face the challenge. As the seismic geology generally have a good effect, the method has been widely used in almost all areas of geological exploration work and stages. The data of seismic is needed to be divided into different classes or clusters and analyzed, so further we can reduce mining costs, find new oil formation, and provide a decision on scientific basis.The K-means algorithm is suitable for massive data clustering; the algorithm is used to cluster the data of geological exploration seismic. But the algorithm has its inherent shortcomings, the improvement of K-means algorithm is needed to cluster the data of seismic.The purpose of this paper is to improve the traditional K-means algorithm, and make it suitable for the analysis on history data of geological exploration seismic. It is verified that the improvement of K-means algorithm is more effectiveness and suitness.The main work completed is as follows:Firstly, the paper gives an overview of the basic theory of data mining techniques, describes the basic principles of cluster analysis, and introduces the major clustering algorithms.Secondly, it analysis’s the limitations of the K-means algorithm and describes the existing approaches of improvement. Further more, it introduces the methods of Kruskal minimum spanning tree algorithm, and improves the traditional algorithm with it.Finally, the actual data is used to verify the both the traditional algorithm and the improvement algorithm. It improves that the improvement algorithm has more accuracy and interpretability.The geological exploration seismic data clustering method will be further improved with the development of data mining.