How ANOMALY DETECTION works in time series using the Holt-Winters Algorithm

This video explains how anomalies are detected in a time-series graph.
The algorithm's name is Holt-Winters. The idea is simple, and the results are often useful.
Cheers!
00:00 What is a time series?
00:30 How can we find anomalies?
01:00 Anomaly Detection Algorithm
02:25 Repeated Differentiation
03:36 Example
04:14 Thank you!
#anomalydetection #timeseries #monitoring
InterviewReady: interviewready.io/
Designing Data Intensive Applications Book: amzn.to/3SyNAOy
Github: github.com/InterviewReady/sys...
LinkedIn: / interview-ready
Twitter: / gkcs_
#SystemDesign #InterviewReady #Coding

[@GautamKhatter]

Amazing how high school mathematics is solving real business problems.

[@asrjy]

Great video! I have a question. Why does 25-34 in the third graph show anomalies but in the original graph it seems normal?

[@prasenlonikar9753]

Gaurav, you have very well explained here.
I have a question - let's say i want to predict anomalies. I know how to detect them,
How would you do it ?

[@gkcs]

Predicting is a bit impossible, no? It won't be an "anomaly" if it is an "expected event".

[@mohinicoder]

I did this in my 2nd job at startup. Using derivative we were identifying spikes

[@KomalYadav-di3qz]

Easy peasy lemon squeeze. Thanks

[@J0Y22]

So how you will imply this?

[@RishiRajxtrim]

👍🙏💯