TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Simple anomaly detection for metrics with a weekly pattern

42 pointsby iliasflover 9 years ago

3 comments

jhundover 9 years ago
I&#x27;m wondering if there is added value in replacing the proposed simple alarm function with control charts: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Control_chart" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Control_chart</a>
评论 #11054965 未加载
SixSigmaover 9 years ago
In Six Sigma we track stability using what is known as &quot;The 7 Run Rule&quot; [1]<p>Not only do you want to see spikes but also unusual activity that suggests further investigation is warranted.<p>[1] <a href="https:&#x2F;&#x2F;www.qimacros.com&#x2F;control-chart&#x2F;stability-analysis-control-chart-rules&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.qimacros.com&#x2F;control-chart&#x2F;stability-analysis-co...</a>
d0ugieover 9 years ago
I&#x27;m looking to use machine learning somehow to get a baseline of weekly listener activity of the scanner feeds on broadcastify in order to spot anomalies of potential newsworthy events. This is motivation that what I have in mind is doable.