Social network public opinion situation monitoring method based on pre-training model

• Main Authors: LIU QIANG, JI YIMU, ZHU JINSEN, LIU SHANGDONG
• Format: Patent
• Language: Chinese; English
• Published: 01.12.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention discloses a social network public opinion situation monitoring method based on a pre-training model, and belongs to the technical field of text information mining. Constructing an LDA model to realize event clustering on the text data; performing word granularity-based coding on the text data by using the pre-training model after LoRA fine tuning, keeping the text coding length consistency through filling or truncation operation, and outputting an emotion classification result through a recurrent neural network and a full-connection neural network; according to an event clustering result and a text sentiment classification result, respectively obtaining attention change and sentiment change, and meanwhile, realizing situation prediction based on the attention change and the sentiment change; event clustering and text sentiment classification are adopted to realize sentiment analysis of events in the time dimension, so that real-time monitoring of public opinion situations of different events is