The Research On Diffusion Law And Early Prediction Of Network Events

With the coming of4G high-speed network era, the propagation scheme of NetworkEvents, which is real-time in interaction and low in cost, has different degrees of effects onsociety lifes, policies and regulations and the mentalities of the people. So it brings aboutproblems on monitoring and leading of the public opinion.My article carries out thereal-time monitoring and evaluation for the early characteristics of network events, andpredicts its future development, in order that we can alert and guide a healthy developmentof the public opinion.My article studies the diffusion model of network events, I collected the earlycharacteristics and the final diffusion range of a lot of network events, I recorded a fullevaluation and its contents, I summarized the time regular, the heat regular and thebiological clock regular of network events diffusion, and rule of information filtering fordifferent network events, rule of information filtering for the same network event.We established a network event propagation model based on the diffusion regular ofnetwork events. We established the logical relationships between data of network event andevaluated the influence of network events, by classifying the real-time network events anddetecting the diffusion and tendentiousness; we forcasted the scale and the direction of thedevelopment of real-time network events, we decided the warning level, moreover, thedirection and measures to the public opinion, we guided the network events at the initialstage, as a result we can control the development and the influence within the reasonablescope.We designed a self-adjusting predictive model based on the Genetic Algorithm andestablished a system on the background of big data which is calculated simply,datarandomly,proceed fast. We forcasted the development of the network events occurring atthe same time and acquired a good result, then we proved its useful value and itspromotional value.Innovation points:1. We find that the network events increase rapidly in gold24hours.The developmentof real-time network events predicted system.Network events early to be controlled.2. Network Event prediction model based on genetic algorithm.Networking eventdiffusion model Self-Learning.It has a simple calculation, faster processing time, adaptiveadvantages.