PROJECT TITLE :

User Behavior Prediction of Social Hotspots Based on Multi message Interaction and Neural Network

ABSTRACT:

The diversity of messages under social hot subjects plays a significant influence in user engagement behavior in network public-opinion analysis. This study provides a prediction model of user participation behavior during multiple messaging of hot social subjects, taking into account the interplay between numerous messages and complicated user behaviors. To begin, a multimessage interaction influence-driving mechanism was presented to better forecast user participation behavior by taking into account the influence of multimessage interaction. Second, this study proposes a user participant behavior prediction model of social hotspots based on a multimessage interaction-driving mechanism and the BP neural network, in light of the behavioral complexity of users engaging in multimessage hotspots and the simple structure of backpropagation (BP) neural networks (which can map complex nonlinear relationships). Finally, the multimessage interaction has an iterative directing effect on user behavior, making the BP neural network easily overfit. To avoid this issue, a simulated annealing approach is used to increase the prediction accuracy of the standard BP neural network. The model not only predicted user participation behavior in actual circumstances of multimessage interaction in evaluation studies, but it also measured the relationships between many messages on hot themes.