基于贝叶斯模型的微博网络水军识别算法研究

发布时间：2018-05-13 01:25

本文选题：网络水军 + 水军识别　；参考：《通信学报》2017年01期

【摘要】：为了能够有效地识别水军,在以往相关研究基础上,设置粉丝关注比、平均发布微博数、互相关注数、综合质量评价、收藏数和阳光信用这6个特征属性来设计微博水军识别分类器,并基于贝叶斯模型和遗传智能优化算法实现了水军识别算法。利用新浪微博真实数据对算法性能进行了验证,实验结果表明,提出的贝叶斯水军识别算法能够在不牺牲非水军识别率的情况下,保证水军识别的准确率,而且提出的阈值优化算法能显著提升水军识别的准确率。
[Abstract]:In order to be able to identify the navy effectively, based on the previous relevant research, we set the fan concern ratio, average Weibo number, mutual concern number, comprehensive quality evaluation. The six characteristic attributes of collection number and sunshine credit are used to design Weibo water army recognition classifier, and based on Bayesian model and genetic intelligence optimization algorithm, the recognition algorithm of water army is realized. The performance of the algorithm is verified by using the real data of Sina Weibo. The experimental results show that the proposed Bayesian recognition algorithm can ensure the accuracy of water army recognition without sacrificing the recognition rate of non-water army. Moreover, the proposed threshold optimization algorithm can significantly improve the accuracy of water army recognition.
【作者单位】：中央财经大学信息学院;电子科技大学网络与数据安全四川省重点实验室;新疆财经大学计算机科学与工程学院;
【基金】：国家自然科学基金资助项目(No.61602536,No.61273293,No.61309029) 北京市社会科学重点基金资助项目(No.16YJA001) 网络与数据安全四川省重点实验室开放课题基金资助项目(No.NDSMS201605) 中央财经大学学科建设基金资助项目~~
【分类号】：TP18

【相似文献】