Given the wide range of detection methods available, it is important to evaluate their performance to determine their ability to detect events and avoid missing significant ones correctly. Accurately assessing whether detection occurs close to actual events requires introducing time tolerance in event detection evaluation. Section 7.1 outlines the general metrics used for point events. Section 7.2 introduces the concept of time tolerance. Section 7.3 discusses interval-based metrics. Section 7.4 presents novel metrics for online event detection. Section 7.5 describes representative benchmarks for event detection. Finally, Sect. 7.6 briefly highlights advanced topics in the field.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Evaluation Metrics

  • Eduardo Ogasawara,
  • Rebecca Salles,
  • Fabio Porto,
  • Esther Pacitti

摘要

Given the wide range of detection methods available, it is important to evaluate their performance to determine their ability to detect events and avoid missing significant ones correctly. Accurately assessing whether detection occurs close to actual events requires introducing time tolerance in event detection evaluation. Section 7.1 outlines the general metrics used for point events. Section 7.2 introduces the concept of time tolerance. Section 7.3 discusses interval-based metrics. Section 7.4 presents novel metrics for online event detection. Section 7.5 describes representative benchmarks for event detection. Finally, Sect. 7.6 briefly highlights advanced topics in the field.