摘要
以提升网络热门舆情分类准确率,降低分类时间为目标,提出了基于数据挖掘技术的网络热门舆情分类方法。将小波核函数和支持向量机结合构成小波模糊支持向量机,采用增量学习机制和贝叶斯分类算法建立增量贝叶斯分类算法,组成小波模糊支持向量机-增量贝叶斯分类算法解决测试样本易分类失误以及类条件独立假定性很难获取问题,通过计算待测样本和小波支持向量机之间的距离,实现网络热门舆情分类。经实验验证:类置信度较高时,文中方法分类准确率高,运行时间少,可快速分类网络热门舆情,且网络热门舆情分类结果的查全率以及查准率都在94%以上,分类精度较好。
In order to improve the classification accuracy of network popular public opinion and reduce the classification time,this paper proposes the classification method of network popular public opinion based on data mining technology.The wavelet fuzzy support vector machine(wfsvm)is improved by using wavelet kernel function and support vector machines.The incremental Bayesian classification algorithm is obtained by using incremental learning mechanism which further forms the wavelet fuzzy support vector machine(wfsvm)-incremental Bayesian classification algorithm.This could solve the problems of easy classification error of test sample and difficulty of acquisition of class condition independence assumption.The distance between wavelet support vector machines is used to realize the network popular public opinion classification.The experiment results show that when the class confidence is high,the classification accuracy of the research method is high,and the running time of different test samples is low,which can quickly classify the network popular public opinion.Besides,the recall and precision of the network popular public opinion classification results are more than 94%,and the classification accuracy is good.
作者
杨小艳
YANG Xiao-yan(College of Electronic&Information Engineering,Ankang University,Ankang 725000,Shaanxi Province,China)
出处
《信息技术》
2022年第2期59-63,68,共6页
Information Technology
基金
安康市科技计划项目(2018AK02-12)。
关键词
数据挖掘
网络热门舆情
小波核函数
支持向量机
朴素贝叶斯
data mining
network popular public opinion
wavelet kernel function
support vector machines
naive Bayes
