The Comparison Between BP Artificial Neural Network And Logistic Regression Models In Pattern Classification

Pattern Classification is a very important problem in many fields. And statistical methods and Artifical Neural Network are the most usefull methods in Pattern Classification, so compare and study that two methods is very nessary.Atificial Neural Net-works(ANN) that simulate the biologic the Neural Network are mathematical models, and there are some different aspects between ANN and statistic methods constitutionally, but there are also many homologous aspects between them. Beside comparing the Perceptron and Fisher discrimination method which are used to linear classification, combining the examples of how to classifie poor students in NENU, this paper analyses the differences and homologous aspects between BP neural network and Logistic regression model.