NLP-ADBench
收藏资源简介:
NLP-ADBench是由南加州大学等机构创建的自然语言处理异常检测基准数据集,包含8个经过精心挑选和转换的分类数据集,用于评估NLP异常检测方法。数据集大小各异,涵盖了从新闻到社交媒体评论等多种文本类型,旨在识别文本中的异常模式。数据集的创建过程包括从原始数据集中选择合适的文本源,并通过语义区分来定义异常类别。NLP-ADBench主要应用于网络系统的安全性和可靠性提升,如欺诈检测、内容审核和用户行为分析。
NLP-ADBench is a natural language processing anomaly detection benchmark dataset created by the University of Southern California and other institutions. It comprises 8 carefully curated and transformed classification datasets designed for evaluating NLP anomaly detection methods. With varying sizes, the dataset covers diverse text types ranging from news articles to social media comments, aiming to identify anomalous patterns in text. The creation process of NLP-ADBench includes selecting appropriate text sources from original datasets and defining anomaly categories via semantic differentiation. NLP-ADBench is primarily utilized to enhance the security and reliability of network systems, such as fraud detection, content moderation, and user behavior analysis.




