Data Classification And The Application In The Fencing-load-analysis

With the maturity of database technology and the popularization of the data application in the society, the amount of data that mankind has accumulated is increasing in an incredibly rapid speed. The ocean of information and the huge wave of the data are pouring to human beings. Compared with the ever-developing data management technique, tools for data analysis cannot render the knowledge that the decision-makers are longing for. Thus the unique phenomenon of"rich data, poor knowledge"turns up. In order to settle this contradiction, Data Mining comes into being.Classification is a task of great importance in data mining and thus is extensively applied to commerce at present. The destination of classification is to acquire a classification function or a classification model that can map a data item to a pre-assigned class. There are many techniques for data classification such as the decision tree, Bayesian belief network, association-based classification, genetic algorithms, K-nearest neighbor classifiers, neural network, etc.This paper first introduces the definition of classification, its process and construction. Then it focuses on two frequently-used techniques: the decision trees and artificial neural networks. For the former, there are two common methods to build it: the ID3 and C4.5 algorithms based on the information theory and the CART, SLIQ methods based on the lowest GINI index. This paper describes ID3 and C4.5 algorithms in detail, and also provides some examples for illustration.Artificial neural network (ANN) is a nonlinear dynamic system, which is constructed by theoretical abstracting and predigesting of the biological neural network in terms of structure, function and some basically specific aspects .ANN has distributed storage mode and parallel disposing mode of information.Back-propagation network is representative network model in these different network models, which has got broad applications.