Multidimensional Information Mining System To Achieve Its Application

Data mining play a very key role in the course of logging well's information processing. And at present, typical data mining systems have weak interaction, have low controllability in the course of data mining and are poor in data visualizing more or less. For the actual need of logging well's information processing, according to the scientific procedure of data mining, we have developed the Multiple Dimensions Information Mining System. In this system, we have utilized the technique of computer simulation—visual programming tool VC++, the technique of database management and the technique of data visualizing. This system has included several data mining methods, such as BP neural network, genetic algorithm, SOM network and TFI clustering, etc. This article has also put forward the key procedure and method how to set up high-precision and high-efficiency classification model, regression model and clustering model with the system.The major achievements obtained are listed as follows:1 An independent system for multiple dimensions information mining system has been developed. It doesn't matter to great data mining systems. It is used in geophysics. It has the input interface to receive logging well's data and grid data, so that the data input course is simple in the data mining course for logging well's data and grid data. It also has the input interface to receive text's data and Excel's data, so that it's possible for other information's data mining.2 The function of scatter points analysis has been developed based on the technique of data visualizing. So that it's convenient to find the relativity of the different data attributes. Several data preprocessing functions have been developed as well.They are aggregation, attribute transformation, odd data marking and sampling. Aggregation as well as attribute transformation can make the data attributes more fitting to different data mining methods. Odd data marking as well as sampling can strengthen the accuracy and adaptability of data mining model.3 Several analytical functions have been developed.They are BP neural network, GA-BP neural network, SOM network and TFI clustering. They have strong interaction, high controllability in the course of data mining, high efficiency and accurate result. They contain three data mining tasks. The total number of data mining...