Public opinion monitoring method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a public opinion monitoring method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a pre-training model which is a language model obtained through the training of a preset network structure through linguistic text data for a language model task; modifying an output layer of the pre-trained model into a text classification layer to obtain an initial model; for the text classification task, training the initial model by utilizing the public opinion sample data to obtain a public opinion monitoring model; and monitoring the obtained public opinion data based on the public opinion monitoring model. In the embodiment of the invention, the initial model is directly constructed on the basis of the pre-training model, i.e., the parameters of the pre-training model are reserved in the initial model, so that the model does not need to learn from zero, and the model training efficiency and the model prediction accuracy are improved; and thus, the accuracy of detecting the public opinions by utilizing the trained public opinion monitoring model is ensured.

Description

technical field [0001] The present application relates to the technical field of the Internet, and in particular to a public opinion monitoring method, device, electronic equipment and storage medium. Background technique [0002] With the development of the Internet, online media has penetrated into people's daily life, and more and more users choose to vent their emotions on social media platforms, so how can companies understand the positive and negative public opinions of their own companies in the first place? , and how to quickly grasp the types of problems that negative public opinion belongs to when negative public opinion appears is a very important part to ensure that enterprises can quickly launch public relations emergency strategies in the first place, so as to improve service quality and customer satisfaction. [0003] The existing network public opinion monitoring and analysis methods are mainly divided into manual monitoring and technical monitoring. Manual ...

AI Technical Summary

Problems solved by technology

Manual monitoring is a traditional information monitoring method, mainly with the help of some free public opinion monitoring tools or search engines, site navigation search of news portals, etc. The disadvantage is that the network information is always changing dynamically, and the amount of information is huge. It is difficult to guarantee the timeliness of information and the comprehensiveness of acquisition; most technology monitoring methods use machine system programming search or traditional artificial feature engineering + machine learning methods of shallow classification models. This traditional machine learning method text representation is high latitude and high Sparse, the ability to express features is very weak, and the neural network is not good at processing such data, it is difficult to solve inaccurate word segmentation, lack of standard and complete emotional lexicon, negative word problems, problems in different scenarios and fields, etc., making public opinion The accuracy of monitoring is low

Images