改进潜在语义分析和支持向量机算法用于突发安全事件舆情预警

发布时间：2018-02-13 17:24

本文关键词： 潜在语义分析支持向量机舆情预警情感倾向性分析　出处：《数据分析与知识发现》2017年02期 　论文类型：期刊论文

【摘要】：【目的】针对现有预警体系多以企业自身和监管部门为主体、忽视网络舆情,导致预警力度不强、缺乏透明度及敏感性、使突发性安全问题时有发生且无法得到及时处理的现状,提出一种新的舆情预警模型。【方法】通过元搜索技术挖掘舆情信息,增加基准偏移值优化情感特征项倾向性权重,添加修正因子以改进潜在语义分析和支持向量机(LSA+SVM)算法,构建舆情分类预警模型。【结果】以多组突发性安全事件为例,应用Matlab进行仿真实验。结果证明该舆情预警模型切实可行,反应迅速,在语义维度为10时准确率可达85.75%。【局限】此方法对于能引起关注和讨论的安全事件更加有效。【结论】改进算法适用于舆情预警,可为企业和监管部门根据分类结果及时采取有效的预警措施提供合理化建议。
[Abstract]:[Objective] the existing early warning system in enterprises and supervision departments as the main body, ignoring the network public opinion, leading to early warning efforts is not strong, lack of transparency and sensitivity to sudden security problems have occurred and the situation can not get timely treatment, put forward a new public opinion warning model. [method] public opinion information mining through meta search technology, optimize the emotion feature tendency to increase the weight reference offset value to improve the latent semantic analysis and support vector machine (LSA+SVM) algorithm with correction factor, build public opinion classification prediction model. [result] to multiple security incidents as an example, the application of Matlab simulation experiments. The results show that the early warning of public opinion the model is feasible, rapid response, in the semantic dimension is 10, the accuracy of 85.75%. [] limited this method for security events can cause the attention and discussion more effective. [ Conclusion the improved algorithm is suitable for the early warning of public opinion, and it can provide reasonable suggestions for enterprises and supervision departments to take effective early warning measures according to the classification results.

【作者单位】：哈尔滨理工大学管理学院;
【基金】：国家自然科学基金项目“高技术虚拟产业集群运行模式研究”(项目编号:70873029) 黑龙江省自然科学基金项目“高新技术企业物流模式选择技术研究”(项目编号:G201203) 黑龙江省博士后科研启动资金资助项目“黑龙江省制造企业动态联盟信誉保障机制研究”(项目编号:LBH-Q12065)的研究成果之一
【分类号】：TP18

【相似文献】