The Analysis And Research Of Regress-Based Prediction System On Data Mining

Main research of this paper is to solve the problem that how to settlethe existed data record effectively to derive automatically the spreadingdescribe of the definite data, thus which can predict the data in thefuture. Regression, Decision tree , Neural Network can be used inpredicting .The difference between the method of Decision Tree and NeuralNetwork is that the output of Regression is continuous number value, butthe output of decision tree and Neural Network are the dispersedclassification value. We will discuss mainly that the method of RegressionAnalysis is applied in the Data Mining predicting system.Regression Analysis is one of the important methods in the Data Miningpredicting system, and it can be divided into Linear Regression andNon-linear Regression. And the optimization criterion of estimation forthe estimated parameter is usually method of least squares during theRegression Analysis. This paper introduced four kinds of optimizationalgorithm of Regression Analysis: method of least squares,Neural Networkfuzzy perceptron method,Neural Network BP algorithms Another kind isgenetic algorithms based regression analysis regression analysis.Alsosummarized various kinds of implementation methods in the predictingsystem, and have compared them.Finally, this paper has applied some kinds of commonly used RegressionAnalysis algorithms in the real Data Mining predicting system. It explainfatherly Regression Analysis algorithm is extremely important in DataMining through the analysis of real data, and can get up to very goodprediction result in various kinds of data analysis.