Anomally score in temperature value series

#1

Hi, I have a certain application for stream data anomally detection and would like to know if a suitable approach is possible with Edge Impulse.

I attach my dataset. It is a time series with temperature values. Each value has a label 0 or 1 (0-no anomally 1-anomally). I would like to train a model with such a dataset and estimate an anomally score for next values. https://github.com/Telepresent/Sigze/blob/main/Data_buenoSigze.csv

Could you please tell me if there is a suitable approach for something like this in Edge Impulse?

#2

Hello @iarakis ,

Yes that should be doable.
You could also train a regression model to predict the future values: https://www.edgeimpulse.com/blog/predict-the-future-with-regression-models

Regression documentation: Regression (Keras) - Edge Impulse Documentation

Anomaly detection documentation: https://docs.edgeimpulse.com/docs/edge-impulse-studio/learning-blocks/anomaly-detection

Regards,

Louis