Research On Summarization Methods For Public Opinion News Texts Involved In Cases

With the rapid development of the Internet,the production and dissemination of case-related public opinion information has become unprecedentedly rapid.Due to the sensitivity and particularity of case-related public opinion,higher requirements have been imposed on the monitoring of the case-related public opinion.The automatic text summarization technology can quickly extract an important piece of information that is far less than the original data amount from a large number of news texts.It can help users quickly grasp relevant public opinion information and has an important role in public opinion monitoring.This study focuses on the characteristics of the case-related public opinion,analyzes the role of the case elements in the abstract extraction process,and researches the single document summarization and the multi-document summarization.Mainly completed the following research work:(1)Method for obtaining public opinion news data and corpus constructionData acquisition and labeled datasets are the basis of supervised learning.This study is based on the scrapy crawler framework,using a non-template web content extraction method,and using keyword searches to collect public opinion news data from Baidu News and Sogou News.According to the different task requirements of single-document summarization and multi-document summarization,two sentence tagging methods are proposed to construct the data sets respectively.It has an important supporting role for the later research points.(2)Text summary method involving the sensation of the case based on the case ele-ment guidanceCompared with the task of text summarization in open domain,the summary of public opinion usually involves specific case elements that are of great guiding effect in the process of summary generation.Therefore,a case-related news text summarization method is proposed based on depth learning framework.First,the case elements are defined.Then,through the attention mechanism,the case element information is integrated into the dou-ble-layer coding process of words and sentences in the news text to generate the news text representation that contains the case element information.Finally,the multi-feature classification layer is used to classify the sentences.In order to verify the validity of the algorithm,experiments are conducted on the public opinion news summary data set.The experimental results show that the proposed method has better performance than the base model,and is more effective and advanced.(3)Method for summarizing texts of case-related public opinion news based on case element guidanceIn order to better grasp the case-related public opinion,only a single document summarization is not enough,but also a multi-document summarization corresponding to different cases.In text clusters,sentences and sentences are associated with each other,case elements and sentences also have varying degrees of association.These associations play an important role in extracting abstract sentences.A case-related public opinion summarization method based on graph convolution of sentence association graph with case elements is proposed,which uses graph structure to model all text clusters,with sentences as the main node,words and case elements as auxiliary nodes to enhance the relationship between sentences.uses multiple features to calculate the relationship between different nodes.Then use graph convolutional neural network to learn this sentence association graphs,and classify the sentence to obtain the candidate summary sentence.Finally,the sentence is deduplicated and ranked to obtain the case-related public opinion summarization.Experiments are performed on the case-related public opinion summary dataset.The experimental results show that the method achieves better results than the benchmark model,indicating that both the composition method and the graph convolution learning method are effective.(4)Prototype system of case-related news summarization.Based on the above research results,a prototype system of public opinion news summarization was designed and implemented.Integrated data acquisition,single document summarization model and multi-document summarization model.Provides users with a visual information acquisition platform.