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

Problem about the dataset dimensions #43

Closed KcAcoZhang closed 3 years ago

KcAcoZhang commented 4 years ago

Dear Author,

As you said, the data in column(dimension)0 is real telemetry data, so what role does the data from other dimensions play in this project, and do they influence the results(performance)? Can I keep only the 0 th dimension of real telemetry data for my training?

khundman commented 3 years ago

The other dimensions represent one hot-encoded command data, which affects telemetry behavior. We didn't do experiments with only the telemetry channels in the 0th dimension, but it would be interesting to see that comparison.