摘要
针对物流新闻类别分布不均衡,分类器容易倾向大类别而忽略小类别的问题,提出基于加权补集的朴素贝叶斯分类模型。构建物流新闻语料库,结合卡方检验进行特征选择,基于局部、全局和类内、类间的思想,分析并改进传统特征加权算法,设计适用于类别分布不均衡物流新闻的加权补集朴素贝叶斯模型。实验结果表明,相较传统分类方法,加权补集朴素贝叶斯模型能有效解决物流新闻文本不均衡情况下的分类问题,快速准确地对物流新闻进行分类。
Focusing on the problem of the imbalanced distribution of logistics news categories and the tendency of classifier to favor large categories but ignore small ones,a naive Bayesian classification model based on weighted complements was proposed.A logistics news corpus was constructed,chi-square test was carried out for feature selection and the traditional feature weighting algorithm was improved based on the ideas of local,global,intra-category and inter-category to design a weighted complement naive Bayesian model for logistics news with imbalanced category distribution.Experimental results show that compared with the traditional methods,the proposed model can effectively solve the classification problem caused by the imbalance of logistics news and classify logistics news quickly and accurately.
作者
许英姿
任俊玲
XU Ying-zi;REN Jun-ling(School of Information Management,Beijing Information Science and Technology University,Beijing 100192,China)
出处
《计算机工程与设计》
北大核心
2022年第1期179-185,共7页
Computer Engineering and Design
基金
国家重点研发计划基金项目(2019YFB1405003)。
关键词
朴素贝叶斯
不均衡样本
补集
物流新闻
文本分类
特征加权
Naive Bayes
imbalanced sample
complement
logistics news
text classification
feature weighting
