Bad data identification method based on clustering analysis coupled neural network prediction

The invention provides a bad data identification method based on clustering analysis coupled neural network prediction, and the method mainly comprises the steps: obtaining original data, carrying out the dimensionless processing, and recording the original data as a sample set; randomly selecting K data samples as K clustering centers; and calculating the distance between each remaining sample and the K clustering centers, and classifying the remaining samples into the class represented by the clustering center with the minimum distance from the remaining samples until all samples in the sample set are classified, thereby obtaining K clustering sample sets. And a BP neural network model is established, aiming at the above classification, a cross validation method is adopted to test each category by using a neural network, and finally a bad data set is obtained. Through the coupling clustering algorithm and the neural network model, the method is very suitable for processing denitration system data with the c