Design And Research Of Intelligent Question Answering System Based On Chinese Community

In recent years,with the rapid development of mobile Internet and e-commerce,there is a huge amount of people keen on online shopping,using the Internet to express their opinions and publishing their views on goods and services on the Internet.In the shopping platform,on one hand,analyzing these comments can not only help the merchants to grasp the advantages and disadvantages of the products in time,but also improve the quality of them,which contributes a lot to satisfy the needs of consumers and increase the scale of the business.On the other hand,it also provides a reference for consumers to have a comprehensive understanding of the product information.Therefore,how to use the existing natural language processing technology to process and analyze the emotional tendency of these texts,which has become one of the active fields many researchers focused on.Traditional sentiment analysis methods rely on large emotional dictionaries and complex Feature Extraction Engineering.If the sentiment dictionary is not perfect,the emotional words in the comments will not appear in the sentiment dictionary,and the emotional polarity will be unable to judge,either.As the increasing scale of the text data,manual data labeling needs too much human labor and material resources,and certain domain knowledge.which limits the development of these methods.In this paper,the deep learning theory and natural language processing technology are combined,on the basis of text classification method based on convolutional neural network,a textual sentiment analysis based on piecewise convolutional neural network integrated with feature fusion is proposed.This model is able to extract the main feature of sentences in segments;and utilize the method of part of speech and word vector fusion to solve the problem that the word vector cannot distinguish the synonym.As is shown by the experimental results,the proposed method can remarkably improve the precision,recall rate of emotional analysis tasks compared with the traditional text convolution neural network,and consequently,a sentiment analysis system for product reviews based on the above algorithm was implemented.