科技回声

I'm not sure why an ad that only tells its readership "it's too complex, just use our product" was posted here.While I'm new to this field myself, the standard score[1] and interquartile range[2] algorithms can be utilized for this purpose, where the underlying calculations use a "streaming" algorithm eliminating the need for storage, such as [3].I recently used some of this to monitor abnormally high request times for an application at my workplace.[1] <a href="https://en.wikipedia.org/wiki/Standard_score" rel="nofollow noreferrer">https://en.wikipedia.org/wiki/Standard_score</a>[2] <a href="https://en.wikipedia.org/wiki/Interquartile_range" rel="nofollow noreferrer">https://en.wikipedia.org/wiki/Interquartile_range</a>[3] <a href="https://nestedsoftware.com/2018/03/27/calculating-standard-deviation-on-streaming-data-253l.23919.html" rel="nofollow noreferrer">https://nestedsoftware.com/2018/03/27/calculating-standard-d...</a>

I’ve used Inter-Quartile Range for detecting anomalies and found it pretty decent and without needing tons of dependencies or ML, just some C#.How does it compare to other methods?<a href="https://www.yourdatateacher.com/2021/11/01/outlier-identification-using-interquartile-range/" rel="nofollow noreferrer">https://www.yourdatateacher.com/2021/11/01/outlier-identific...</a>

Anomaly Detection for Time Series Data: An Introduction

2 条评论

Anomaly Detection for Time Series Data: An Introduction

2 条评论