Based Adapter Grid Clustering Arithmetic And The Study Of Application In The Information Extraction

Data Mining is a tool for people finding information and knowledge from huge data. Clustering can find rules behind the huge data, it can also combine with other data mining methods for finding more knowledge, and advance the efficiency and quality of data mining, it is an important research part of data mining. So investigation of the clustering's arithmetic has important meaning in theory and practice. How to quickly clustering and how to get better result of clustering are the emphasis and difficulty of clustering arithmetic research. Based density method can find any kind of class but the running time is so long and it can't find the class which has different data distribute. While based grid method have high running speed, but it reduce the quality of the result. The paper based on the researching of these methods, combining with the excellence of the grid and density methods, applying the adapter gird to grid partition. It improves the quality of grid partition by refining the edge of the class, so improves the quality of the clustering. Because the query analysis is a common operation, and the database updating usually, so the paper also research how to query and incremental clustering. The experiment results show the arithmetic is correct and has high precision. Later the paper apply the grid partition method to CLIQUE in finding density regions of 1-dimension, this method can find reasonable density region and reduce the amount of the candidate density region, so improve the veracity of clustering. Now remote sensing images have became the huge data source of Geographic Information System. Auto clustering and information extraction of the remote sensing images are the key of remote sensing disposal. Now we are in the vigorous phase of this part research. This paper discuss the application of clustering in remote sensing images classify and information extraction, using the spectral character, and applying the based adapter clustering to classify and information extraction, obtaining the primary application production.