News text-oriented event line extraction method based on deep clustering model

Abstract

The invention discloses a news text-oriented event line extraction method based on a deep clustering model. The method comprises the steps of preprocessing news texts; wherein an initial hidden eventof the text represents pre-training; grouping news texts in the corpus according to the release time; determining an event line to which each piece of news in each group belongs based on a deep clustering model; performing post-processing on the event elements with the same event line number in each group to obtain structured display of the event; and carrying out post-processing on each group ofextracted events with the same event line number to obtain an event line. The neural network model is adopted to automatically extract the event features implicit in the text, manual feature selectionand construction are avoided, and the extracted implicit event features of the text can provide support for downstream applications; event feature extraction and event line extraction can be carriedout at the same time, and the possibility of error propagation is reduced. Compared with a conventional event line extraction method, the method is higher in extraction accuracy and recall rate.

Description

technical field [0001] The invention relates to a method for unsupervised event line extraction of news text by using a computer, and belongs to the technical field of information processing. Background technique [0002] With the rapid development of online news media websites and mobile news applications, the massive news reports generated by social media every day have become the main way for people to obtain and pay attention to domestic and foreign events, which has had a huge impact on society. However, the value of massive news reports varies from high to low, and different people pay different attention to them. Moreover, for some events that last for a long time, people tend to ignore the correlation and development trend between events. Therefore, people urgently need a tool that can automatically extract hot events from massive news texts and show how events change dynamically over time. [0003] Event line extraction mainly focuses on extracting popular time fro...

AI Technical Summary

Problems solved by technology

[0004] Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides an unsupervised event line extraction method based on a deep clustering model for news texts, which can solve the problem that the event line cannot be extracted from the text during the event line extraction process. Containing the defects of event representation, realize the extraction of event lines in news texts without labeling data

Images