The Simulation of Abnormal Data of Accurate Monitoring Algorithm of Dynamic Network Public Opinion
At present,the forms of network public opinion information are diverse,including texts,images and vid-eos.The complexity is high.Network public opinion information contains a lot of rumors,malicious attacks,and false information.How to remove valueless and false information from numerous information is one of the difficulties in dy-namic network public opinion monitoring.In order to address this issue,based on an improved Bayesian algorithm,a method of monitoring abnormal data in dynamic network public opinion was presented.Firstly,dynamic network public opinion data were collected through web crawlers,and then the noise was filtered out through an improved multi-wavelet transform coefficient algorithm.Secondly,the key features of abnormal data in dynamic network public opinion were extracted by the TF-IDF algorithm.Meanwhile,high-frequency data was obtained by the LSI algorithm.On this basis,an improved Bayesian algorithm was applied to construct a Bayesian model.Moreover,all features were input to the model.Finally,abnormal data monitoring of dynamic network public opinion was achieved.Experimental results show that the proposed method can accurately obtain THE abnormal data IN dynamic network public opinion,with shorter time.The relative error is less than 0.04％ .
Improved Bayesian algorithmDynamic networkData denoisingAbnormal public opinion dataMoni-tor
NETL
NSTL
万方数据
