---
id: 20260314-T0-09
title: "字符串异常检测算法比较研究发布"
title_en: "String Data Outlier Detection Algorithms Compared"
url: https://ai.daily.yangsir.net/daily/20260314-T0-09
issue_date: 2026-03-14
publish_date: 2026-03-13T04:00:00.000Z
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2603.11049
---

# 字符串异常检测算法比较研究发布

arXiv论文《Comparison of Outlier Detection Algorithms on String Data》首次系统比较字符串数据异常检测算法。该研究评估了6种方法在文本数据上的表现，发现基于语义距离的算法准确率最高达89%。填补了数值异常检测向文本扩展的研究空白，为NLP安全检测提供新工具。

## English Version

**String Data Outlier Detection Algorithms Compared**

The arXiv paper presents the first systematic comparison of outlier detection algorithms for string data. Evaluating six methods on text data, it finds semantic-distance-based algorithms achieve 89% accuracy. The research bridges the gap between numerical and text outlier detection, offering new tools for NLP security.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2603.11049)

**详情页**：https://ai.daily.yangsir.net/daily/20260314-T0-09

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*