一种逻辑模型树算法在网络舆情中的谣言检测

网络舆情中的谣言对社会危害极大,因此有效地检测网络舆情中的谣言已是当务之急.目前,一些单一机器学习算法被相继应用到谣言检测中.针对这些单一机器学习算法在分类上的局限性,将一种融合逻辑回归与决策树的逻辑模型树方法用于谣言检测上.根据舆情分析报告上采集的数据集,实验结果表明:组合模型逻辑模型树的分类预测准确率比已应用到谣言检测的单一机器学习算法明显要高,逻辑模型树是一种有效的谣言检测方法.

A Logistic Model Tree Algorithm for Rumor Detection in Internet Public Opinion

Rumors in Internet public opinion are very harmful to the society,so it has become a top priority to effectively detect rumors in Internet public opinion.Nowadays,some single machine learning algorithms have been applied to the rumor detection.Aiming at the limitation of these single machine learning algorithms in classification,this paper applies a logistic model tree method integrating logistic regression and decision tree to rumor detection.According to the data set collected from exiting public opinion analysis report,the experimental results show that the combined model,logistic model tree has much higher classification prediction accuracy than these single machine learning algorithms applied to rumor detection.Logistic model tree is an effective method for rumor detection.