基于马尔可夫链的舆情热度趋势分析

摘要

Public opinion often has a hidden trend behind its characteristics shown online.A Markov chain based public opin ion analysis method is proposed.In this method, indicator data of public opinion are collected to get the value of time series for the trend.Then the trend changes and the state spaces are analyzed, and the state transition matrix for the final trend is built.Experiments demonstrate that this method can effectively predict the trend of focus public opinion, and be of signifi cance in guiding and controlling the public opinion.%针对舆情热度的波动范围较大,并会呈现某种趋势的特点,提出了一种基于马尔可夫链的舆情热度趋势分析模型.该模型采集相关热点舆情的指标数据,得到热度的时间序列值;分析热度的趋势变化,划分状态空间,构建状态转移矩阵,预测热度的趋势变化区间.实验表明,该方法能有效地预测热点舆情的走势,进而辅助对舆情的引导和控制.