Adaptive Modified neural network based pattern recognition methods

The present invention relates to the field of pattern recognition, specifically a correction pattern recognition method based on adaptive neural network; first method uses the probabilistic neural network model to classify the input training samples, obtained correctly classified samples and misclassified samples ; was then added probabilistic model on the basis of the structure of the neural network input layer, a central layer and the layer excitation, the adaptive correction model build structure of neural network; probabilistic neural network for after misclassified samples, itself as the center point, calculation and other allow class sizes between the radius of the cluster in the same category error sample, in order to achieve batch correction and re-planning decision of the interface classification model, establish adaptive correction neural network model; adaptive last amended neural network model based on the input of the test sample pattern recognition; pattern classification present invention has high accuracy and strong generalization ability of the model, a good classification real-time performance, a bright future and so on.