Accurate and efficient network public opinion detection and early warning method based on multi-level geography

Abstract

The invention discloses an accurate and efficient network public opinion detection and early warning method based on multi-level geography, which comprises the following steps: S1, analyzing the network public opinion information by using a multi-level geographic knowledge base and an entity identification means, locating the place where the public opinion occurs, and labeling the corresponding geographic label; S2, using the machine learning algorithm to optimize the display effect of public opinion information under different conditions. Through the improved multi-level geographic monitoringscheme, the invention can comprehensively provide the public opinion monitoring of the geographic region concerned for the user. At the same time, the user can monitor the hot spots in the hot spotsarea. This not only ensures the comprehensiveness of the data, but also ensures the focused monitoring of key geography, and helps to improve the processing capacity of public opinion events. At the same time, hierarchical management, but also to avoid duplication of work system, to achieve a hierarchical marker, on-demand distribution.

Description

technical field [0001] The invention relates to the technical field of natural language processing applications, in particular to an accurate and efficient network public opinion detection and early warning method based on multi-level geography. Background technique [0002] With the rapid development of the Internet, especially the mobile Internet, network media, as a new form of communication, has replaced traditional media to occupy our daily life by virtue of its fast and convenient attributes. The number of netizens in our country is increasing rapidly, and the degree of participation in the Internet is high. Whether it is a major domestic or international event, it can be quickly disseminated, and views can be expressed and broadcast in spring through the Internet. Everyone can express their opinions and opinions on the Internet and spread ideas; institutions and departments can easily obtain and utilize the opinions of netizens, and at the same time obtain the social ...

AI Technical Summary

Problems solved by technology

The main reason for the poor effect of traditional methods on public opinion recommendation is that users pay attention to the regional nature of online public opinion, and the names of districts, towns, and villages in domestic provinces and cities have duplicate names. After the geographical location is identified, ambiguity needs to be eliminated, but Due to the strict disambiguation method, there are many data omissions; because the current public opinion monitoring is based on the logic of adding a large number of words according to the account, when users pay attention to the specific occurrence of online public opinion, they cannot be displayed in layers