基于改进的粗糙集算法对网络舆情危机指标简化

文章摘要

基于改进的粗糙集算法对网络舆情危机指标简化

Simplification of Network Public Opinion Crisis Index Based on Improved Rough Set Algorithm

投稿时间：2020-09-01 修订日期：2020-09-01

DOI：

中文关键词: 属性约简、粗糙集、RSBRA、KW检验

英文关键词: attribute reduction, rough set, RSBRA,KW test

基金项目:重庆市教委科学技术研究基金项目（KJ130658）;重庆市自然科学基金项目（cstc2019jcyj-msxm0801）;重庆市教委科学技术研究基金项目（KJ1400521）

作者	单位	邮编
朱光婷^*	重庆师范大学	400000

摘要点击次数: 247

全文下载次数: 0

中文摘要:

针对连续属性的网络舆情数据存在冗余和不确定的问题提出了一种改进的粗糙集方法对网络舆情危机指标进行定量筛选。方法：是根据提出的指标体系建立评价指标初始信息表，对于连续属性采用可降低复杂度RSBRA离散化的高效实现算法进行数据预处理，然后基于粗糙集可识别矩阵计算属性重要度的思想对属性进行约简。结论：此改进方法一是通过RSBRA离散化保证了候选断点的最优组合使得粗糙集约简的网络舆情信息表数据更加准确，二是可识别矩阵计算属性重要度的思想对属性进行约简删除网络舆情中没有实际意义的冗余指标，直观简单又降低了计算的复杂度。

英文摘要:

Aiming at the problem that most of the data of network public opinion belong to continuous attribute, and there is still redundancy and uncertainty, rough set can only deal with discrete data, a new rough set model is proposed to quantitatively screen the crisis index of network public opinion. Based on the proposed index system, the initial information table of evaluation index is established. Firstly, the continuous attribute can be preprocessed RSBRA discretized by an efficient algorithm. Conclusion: The KW test is applied to analyze the significance of the final index and verify the rationality of the new rough set model Sex, this method from the actual meaning of indicators to consider, while ensuring objectivity of index screening, intuitive and simple to reduce the complexity of the calculation is also applicable to big data.

View Fulltext 查看/发表评论下载PDF阅读器

关闭