mdsp anomaly-detection Command¶

Usage: mdsp anomaly-detection|ad [options]

train anomaly detection models and detect timeseries anomalies *

Options:
 -m, --mode [template|train|detect] mode [template | train | detect] (default: "train")
 -o, --on [data|asset] on [data | asset] (default: "data")
 -d, --data time series data file (default: "timeseries.mdsp.json")
 -e, --epsilon threshold distance
 -s, --clustersize minimum cluster size
 -a, --algorithm [EUCLIDEAN|MANHATTAN|CHEBYSHEV] distance measure algorithm [EUCLIDEAN | MANHATTAN | CHEBYSHEV]
 -n, --modelname human-friendly name of the model
 -i, --modelid Insights Hub model id
 -i, --assetid Insights Hub asset id
 -n, --aspectname Insights Hub aspect name
 -f, --from begining of the time range
 -u, --to end of the time range
 -k, --passkey passkey
 -y, --retry retry attempts before giving up (default: "3")
 -v, --verbose verbose output
 -h, --help display help for command

 Examples:

 mdsp ad --mode template --data timeseries.data.mdsp.json 
 creates a template for a time series data file
 mdsp ad --mode train --on data --data timeseries.data.mdsp.json --epsilon 0.5 

 trains a model on the timeserie specified in the data file
 mdsp ad --mode detect --on data --data timeseries.data.mdsp.json --modelid 
 detects anomalities of the timeseries in the data file using the model with specified id
 mdsp ad --mode train --on asset --assetid --aspectname Environment --epsilon 0.5
 trains a model on the time series of the aspect "Environment" of the asset with the id 
 mdsp ad --mode detect --on asset --modelid --assetid --aspectname Environment --epsilon 0.5
 detect anomalities of the timeseries on the specified asset and aspect with selected model