Neural network and learning method of neural network

The invention discloses a neural network of the technical field of computers. The neural network is composed of three layers of networks, including an input layer, a hidden layer and an output layer, and full connection is performed between all the layers. The invention also discloses a learning method of the neural network. The learning method of the neural network comprises the following steps that S1: input information is processed layer by layer from the input layer through the hidden layer; S2: the error value of actual output and demand output is computed layer by layer so that the weight value can be adjusted according to the error value; and S3: the learning error of the neural network is enabled to be minimized. Associative memory can be realized by using the simple neural network, connection between the network units is fixed and connection matrix determination is formed by the vector outer product so that repeated practice is unnecessary, the input vector does not need to be preprocessed, directly