A method for summarizing public opinion news texts based on case elements and bigru
A case and text technology, applied in the field of natural language processing, to achieve the effect of improving the F value
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2021-02-05
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The invention relates to a method for summarizing case-related public opinion news texts based on case elements and BiGRU, and belongs to the technical field of natural language processing. Background technique
[0002] Case-related public opinion refers to Internet public opinion related to judicial cases. Compared with general news public opinion, the public opinion involved in the case is sensitive and specific, which can easily induce mass incidents and cause serious adverse social effects. Therefore, it is very important to quickly and accurately obtain public opinion information and control the development trend of public opinion to prevent and resolve public opinion risks. The text summarization task can simplify the length of the news text by extracting core sentences that can summarize the semantics of the text, which can effectively help users obtain key information about public opinion events from a large amount of public opinion data. Gen...
Examples
Embodiment 1
[0063] Embodiment 1: as Figure 1-2 As shown, based on the case elements and BiGRU's public opinion news text summary method, the specific steps of the case-related public opinion news text summary method based on the case elements and BiGRU are as follows:
[0064] Step1. Acquisition of public opinion news texts involved in the case; use crawlers based on the Scrapy framework to collect relevant news from the Internet, and perform manual proofreading to construct a text dataset of public opinion news texts involved in the case. The relevant information of the dataset is shown in Table 2:
[0065] Table 2 Dataset
[0066] number of texts average number of sentences average text length Training set 17434 15.38 776 validation set 1000 16.46 776 test set 1000 14.42 722
[0067] Step2, mark the training data;
[0068] A greedy search method is used to find a sentence combination with the highest score as a summary sentence through the...