ANN-based Data Mining Approach And Its Applications In GIS

Data Mining is a nontrivial process that we can identify the effective, unknown, potentially useful and ultimately apprehensible pattern from databases. Data mining technique is a cross multi-field for researches and applications. So it receives more attention and becomes one of most popular research in recently.Clustering analysis is an important approach of data mining, and an important content of human activity. As a branch of statistics, clustering analysis has a long research history. This article takes a step forward in the algorithms of clustering analysis and discusses two kinds of clustering analysis methods: k-means algorithm merged in density-based, k-means algorithm merged in competitive learning of neural networks. On the base of the discussion, a new integrated clustering analysis algorithm, and an improved integrated clustering analysis algorithm will be presented. Comparing the four methods, the improved integrated clustering analysis algorithm has more valuable in data mining. In the design of ship route, first is to choose an area, in other word to take out of some data, then to preprocess the data, and then to use the improved integrated clustering analysis algorithm to find clustering points, last of all is to get central density point through calculation, connecting with all of the central density points is the ship route.Artificial neural networks are important approaches of data mining, too. ANN can approximate arbitrarily to nonlinear mapping by learning. However, all of the ANN predictive models just can be used to predictthings with one attributor, for example stock forecast. There have not been researches to be found for prediction of things with more attributors. This article presents a new multi-dimension predictive model based on the diagonal recurrent neural networks (PDRNN) with a parallel learning algorithm. This method can be used to predict not only values, but also some points in the multi-dimension space. And also its applications in data mining will be discussed in the paper. Some analysis results show the significant improvement to ship route prediction using the PDRNN algorithm in database of geographic information system (GIS). In addition, this article has made lots of prediction with Matlab language and got satisfying effects.This article gets some valuable results in the data mining of GIS with the improved integrated clustering analysis algorithm and multi-dimension predictive model based on the diagonal recurrent neural networks (PDRNN) with a parallel learning algorithm.Wang Tianzhen( Power Electronics and Electric Drive)Directed by Tang Tianhao...