Online Chinese News Classification System based on Python Reptile and Convolutional Neural Network
Zhe Zhang1, Genyao Zhang1, Chi Zhao1, PeikeWang2
1College of Mathematics and Computer Science, Yan'an University, Yan'an, 716000, China
2College of Computer Engineering, Huaihai Institute of Technology, Huaihai, 222000, China
Abstract: With the advent of the era of big data, the amount of data has shown an exponential growth. A se-ries of news retrieval platforms represented by Netease News have new content all the time. In order to allow users to more directly understand the current hot spots, the system usesTHUCNews data setsfor training, including a total of 14 categories, 830,000 articles. Using TensorFlow to train a character-level convolutional neural network with a single convolutional layer, a 92.24% accuracy rate was achieved on the validation set. Combined with the Python crawler, it eventually achieved automatic crawling and classification of news information, presented to the user in a concise manner.
Keywords: Crawler; Convolutional neural network; Python