Hybrid Algorithm and Forecasting Technology of Network Public Opinion Based on BP Neural Network

摘要

Based on the hybrid algorithm and BP neural network, the forecasting technology of network public opinion was studied. After detailed introduction of the research background and significance, the concept and development of hybrid algorithm, Hybrid algorithm is analyzed in detail. The user network public opinion classification model under the clustering algorithm is established. Then the data mining experiment of Sina Weibo users' network public opinion is conducted. Using an improved BP neural network algorithm, unsupervised user network clustering and evaluation of the results, finally, the user's Internet public opinion is divided into six categories. The characteristic preferences of each user group are analyzed, with recommendations provided for operation recommendation and promotion.