Research And Implementation Of Chinese Interpretive Opinion Sentence Recognition Method Based On CNN

With the emergence of a series of network platforms,a large number of Chinese opinion texts generated by users have been generated on the internet.There is a huge amount of information on the Internet.Users want to be able to understand the cause of the evaluation,and businesses also hope to get more information to help you improve the explanatory comments.Explanatory opinion sentences recognition is in order to meet this demand,but also for further explanation of the excavation to pave the way.This paper is based on Chinese product and service reviews.Specifically,this paper is divided into the following three main areas of work:(1)Explanatory opinion sentences corpus structure:According to the evaluation information in the field of phone and hotel,building two corpuses of explanatory comments.Making detailed annotation rules and analyzing the corpus.Comparing data characteristics in the field of products and services,and analyzing the different effect on the explanatory comments of the two areas.(2)Explanatory opinion sentences identification based on CNN:Explanatory comments word of recognition tasks is identify from the comments contain explanatory comments of the sentence.Using Convolutional neural networks and maximum entropy method for identification,and in the framework of Convolutional neural networks using a separate pool of methods compare experimental results show that CNN can better solve the explanatory comments word recognition problems.(3)Build explanatory opinion sentences recognition system:This thsis uses CNN to train explanatory opinion sentences recognition model,applies to phone and hotel field respectively.Build explanatory opinion sentences recognition system and test the system.The experimental results show that the CNN based Chinese opinion sentences recognition system has better performance.