基于大数据分析方法的微博热点建模与预测

Microblog hotspots modeling and prediction based on big data analysis

微博热点反映一个社会对某一事件的看法,其受到许多因素的影响,具有一定的规律性,同时具有一定的随机性,数据规模庞大,传统方法无法准确、客观描述,微博热点预测错误大,为此设计基于大数据分析方法的微博热点建模与预测方法。首先对微博热点变化特点进行分析,找到引起微博热点预测错误大的原因,然后收集微博热点历史数据,通过聚类分析选择最优样本点组成训练样本,减少数据的规模,最后引入大数据分析方法建立微博热点预测模型,并与其他微博热点预测方法进行对比测试,所提方法的微博热点预测精度超过95%,预测误差远小于当前其他微博热点预测方法,而且建模与预测时间明显减少,加快了微博热点建模与预测效率,具有更高的实际应用价值。

A microblog hotspot modeling and forecasting method based on large data analysis method is designed.The characteristics of microblog hotspot change are analyzed to find out the reasons for the large errors in microblog hotspot prediction.The historical data of microblog hotspots is collected.The optimal sample points are selected by clustering analysis to form training samples and reduce the size of data.The prediction model of microblog hotspots is established by introducing big data analysis method,and is tested and compared with other microblog hotspot forecasting methods.The accuracy of this method is more than 95%,and its prediction error is much less than that of other micro-blog hotspot prediction methods.Moreover,the time of modeling and prediction is obviously reduced,which speeds up the efficiency of microblog hotspot modeling and prediction,and has high practical application value.