| BP (Back Propagation) neural network, that is, back-propagation of error back-propagation algorithm of the learning process, from the dissemination of positive information and error back-propagation of both the process of composition. Input layer of the neurons responsible for receiving input information from the outside world, and passed to the middle layer neurons; middle layer is the internal information processing layer is responsible for information transformation, in accordance with changes in the capacity of information needs, can be designed for the middle layer, single hidden layer many hidden layer or structure; the last hidden layer to output layer of the transmission of neuronal information, after further treatment after the study completed a deal with the process of being spread by the output layer to the outside world deal with the results of the output information. When the actual output and desired output does not enter the back-propagation phase errors. Through the output layer error, according to error gradient descent of the right to fix the value of each floor, to the hidden layer, input layer back propagation layer by layer. Positive cycle of information dissemination and error back-propagation process, the right value at all levels are continuously adjusted process, but also study neural network training process, this process has been carried out to the network output error reduced to an acceptable level, or pre-established determined by the number of study so far.This paper discusses, through the use of BP neural network algorithm, a professional educational programs and overall student achievement based on data in table form, such as simplifying the data, the use of sub-packet approach to the theory of BP neural network applied to the academic direction of identification model, by calculating each In recognition of the professional match modules to determine the direction of our students. |
PDF Full Text Request