Website classification method and system

• Main Authors: A MANTAI, CUI YUANBO, JIN HONG, LIU CHANGYONG, NIE JIAHE, WANG YU, QING RUNDONG, YANG MANZHI
• Format: Patent
• Language: Chinese; English
• Published: 13.04.2018
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention relates to the technical field of networks, and discloses a website classification method and system. The method comprises the steps of: respectively training all kinds of information included in a webpage, so that respective corresponding information matrixes are formed; jointing the formed multiple information matrixes to form an input layer characteristic matrix; extracting the column characteristic vector number and the characteristic vector dimension of the input layer characteristic matrix, so that a convolutional layer is formed; extracting a column vector maximum obtainedby convolution, performing pooling processing, and obtaining row vectors, the number of which is the same as the convolution kernel dimension, so that a pooling layer is obtained; adding vectors output by the pooling layer into a softmax function, so that a fully connected layer is formed. By means of the website classification method and system disclosed by the invention, the existing website classification accuracy is