search

khundman / telemanom

A framework for using LSTMs to detect anomalies in multivariate time series data. Includes spacecraft anomaly data and experiments from the Mars Science Laboratory and SMAP missions.
https://arxiv.org/abs/1802.04431
Other
1.02k stars 248 forks source link

co-relation #52

Closed bappctl closed 3 years ago

bappctl commented 3 years ago

How the multi variate co-relation between channels achieved? Thanks

khundman commented 3 years ago

This is described in the depth in the paper. If you have more detailed questions or if there are parts that aren't clear please reference them specifically.

bappctl commented 3 years ago

thank you. In labeled_anomalies.csv

a) was below feature in a channel labeled as 'contextual' in reference to some other feature?

chan_id | spacecraft | anomaly_sequences | class | num_values | P-1 | SMAP | [[2149, 2349], [4536, 4844], [3539, 3779]] | [contextual, contextual, contextual] | 8505

khundman commented 3 years ago

No, contextual refers to other nearby P-1 channel values. Please see the definition of a "contextual" anomaly in first paragraph of section 2 in the paper:

"...contextual anomalies are single values that do not fall within low-density regions yet are anomalous with regard to local values."