热点特征深挖下的高效微博热门话题预测  

Micro Blogging Hot Topic Prediction Based on Hotspot Feature with Data Mining

作　　者：谷保平[1] 史丽燕[1] 

机构地区：[1]河南广播电视大学,郑州450046

出　　处：《科技通报》2014年第4期80-82,共3页Bulletin of Science and Technology

基　　金：2013河南省社科联资助项目(SKL-2013-598)

摘　　要：提出一种基于数据挖掘的微博热门话题预测方法,在对微博词汇进行基于词频的热门分类基础上,通过热点发现、特征提取、发现学术领袖、热点追踪、关注学术领袖和热点分析6个阶段对预测进行分析和处理。通过统计的方法实现热门预测结果输出。采用一组网络词汇进行实际的热门预测仿真分析,结果显示,基于数据挖掘的微博热门话题预测方法能够更好的实现微博热门话题的预测,预测结果聚类特性优于传统预测方法,算法收敛特性好,具有很好的预测使用价值。A hot topic prediction method of micro blogging based on data mining was proposed, the word frequency wasused to classify the types firstly. Then the hotspot, feature extraction, academic leaders found, hot track, academic leadersconcern and hotspot analysis were all considered to achieve analysis and processing, finally, the result was output throughstatistical. A serious of network vocabulary was used to test the ability between the improved algorithm and traditional algorithm, the result shows that the algorithm of hot topic prediction with micro blogging based on data mining has a better clustering ability, the convergence of algorithm is good and it has a good value for application.

关 键 词：数据挖掘 微博 热门话题预测 聚类性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]