摘要
热点话题具有突发性和实时性,负面热点话题会对社会稳定产生不利影响,为了准确刻画热点话题变化趋势,建立一种布谷鸟搜索CS(cuckoo search)算法优化支持向量机(SVM)的热点话题预测模型(CS-SVM)。首先对网络文本进行聚类预处理,并获取热点话题的时间序列;然后利用SVM对时间序列进行建模,并采用CS算法优化SVM参数;最后通过仿真实验对CS-SVM模型性能进行测试。仿真结果表明,相对于对比预测模型,CS-SVM模型可以准确描述热点话题的变化趋势,提高了热点话题的拟合精度和预测精度。
Hot topics have paroxysmal and real-time properties,negative hot topics can produce adverse effects on social stability. In order to accurately depict the trend of hot topic changes,in this paper we build a hot topic prediction model（ CS-SVM）,it is based on optimising support vector machine（ SVM） with cuckoo search（ CS） algorithm. First it preprocesses the network documents with clustering,and obtains the time sequence of hot topics; then it uses SVM to model the time series,and uses CS algorithm to optimise SVM parameters. Finally,the performance of CS-SVM is tested through simulation experiment. Simulation results show that compared with contrast prediction models,CSSVM model can accurately describe the trend of hot topics change and improves fitting and prediction precisions of hot topics.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第4期330-333,共4页
Computer Applications and Software
关键词
热点话题
布谷鸟搜索算法
支持向量机
文本聚类
Hot topics Cuckoo search algorithm Support vector machine Text clustering
