【摘要】 主要研究网络评论信息的情感倾向性判断的方法,针对于传统IG算法和TF-IDF算法所存在的不足,提出了改进的IG算法和TF-IDF算法。并针对于朴素贝叶斯方法、KNN算法以及SVM分类算法的不足,提出了融合分类器。通过实验表明,该融合分类器取得了一定的效果,能有效的提高分类准确率。更多还原

【Abstract】 This paper mainly studies the method of judging the emotional orientation of network comment information. Aiming at the shortcomings of traditional IG algorithm and TF-IDF algorithm,an improved IG algorithm and TF-IDF algorithm are proposed. Aiming at the shortcomings of Naive Bayesian method,KNN algorithm and SVM classification algorithm,a fusion classifier is proposed. Experiments show that the fusion classifier achieves certain results and can effectively improve the classification accuracy.更多还原