数据挖掘的网络热点话题演化动态预测

Research on Dynamic Prediction of Network Hot Topic Evolution Based on Data Mining Technology

为了及时发现网络热点话题走向,提出基于数据挖掘技术的网络热点话题演化动态预测算法。首先运用基于块的单遍聚类算法,将网络热点话题划分成不同文本集,根据特征划分到不同块中,进行网络热点话题的数据挖掘,然后计算页面话题热度值,利用热度值描述报道获得的先后顺序,判断能否产生新的网络热点话题,实现网络热点话题预测。经实验验证算法网络热点话题动态预测运行时间少;热点话题预测漏检率低,可精准描述热度值变化趋势。

In order to find out the trend of hot topics in the network in time,a dynamic prediction algorithm for the evolution of hot topics based on data mining technology is proposed.First,using the block-based single-time clustering algorithm to divide the network hot topics into different text sets,divide them into different blocks according to the characteristics,and conduct the data mining of the network hot topics.Then,the topic heat value of the page is calculated,the sequence of reports is described by using the heat value,and whether new hot topics can be generated and the prediction of network hot topics can be realized.The experimental results show that the algorithm has less dynamic prediction time for hot topics in network;The prediction of hot topic is low,which can accurately describe the trend of heat value.