Data identification method based on associative clustering deep learning neural network

摘要

To provide a data identification method based on an associative clustering deep learning neutral network.SOLUTION: Firstly, a label set corresponding to an N-class data sample set is acquired and preprocessed, at the same time, a data preset format and a label preset format are acquired, after that, a training is performed to a deep learning neutral network, then, an arbitrary one of test data of each class is set as an input of the deep learning neutral network of the class after converting the test data to the data preset format of the class, a corresponding test output label is acquired, a possible output label and the most preferable output label are defined on the basis of a calculation of similarity between the number of elements of the label set in which a test output label exists, and the data set, after that, a provability that the output labels of the classes coincide with one another or not is calculated, and finally, the possible output label, the most preferable label and the provability are outputted.SELECTED DRAWING: Figure 1