Plugins

TsSTL

@Plugin(type = SparkCompute.PLUGIN_TYPE)
@Name("TsSTL")
@Description("A time series transformation stage to decompose each time signal into seasonality, "
	+ "trend and remainder component leveraging an STL algorithm (Seasonal and Trend decomposition "
	+ "using Loess). This transformation stage adds 'seasonal', 'trend' and 'remainder' fields to "
	+ "the each time record.")
public class TsSTL extends STLCompute {

    ...

}

Parameters

Time Field The name of the field in the input schema that contains the time value.
Value Field The name of the field in the input schema that contains the value.
Group Field The name of the field in the input schema that specifies data groups.
Algorithm Configuration
Outer Iterations The positive number of cycles through the outer loop. More cycles here reduce the affect of outliers. For most situations this can be quite small. Default is 1.
Inner Iterations The positive number of cycles through the inner loop. Number of cycles should be large enough to reach convergence, which is typically only two or three. When multiple outer cycles, the number of inner cycles can be smaller as they do not necessarily help get overall convergence. Default value is 2.
Periodicity The periodicity of the seasonality; should be equal to lag of the autocorrelation function with the highest (positive) correlation.
Seasonal Smoother The length of the seasonal LOESS smoother.
Trend Smoother The length of the trend LOESS smoother.
Level Smoother The length of the level LOESS smoother.

TsAutoSTL

@Plugin(type = SparkCompute.PLUGIN_TYPE)
@Name("TsAutoSTL")
@Description("A time series transformation stage to decompose each time signal into seasonality, "
	+ "trend and remainder component leveraging an STL algorithm (Seasonal and Trend decomposition "
	+ "using Loess). The periodicity required for the SL algorithm is determined automatically "
	+ "through an embedded ACF. This transformation stage adds 'seasonal', 'trend' and 'remainder' "
	+ "fields to the each time record.")
public class TsAutoSTL extends STLCompute {

    ...

}

Parameters

Time Field The name of the field in the input schema that contains the time value.
Value Field The name of the field in the input schema that contains the value.
Group Field The name of the field in the input schema that specifies data groups.
Algorithm Configuration
Outer Iterations The positive number of cycles through the outer loop. More cycles here reduce the affect of outliers. For most situations this can be quite small. Default is 1.
Inner Iterations The positive number of cycles through the inner loop. Number of cycles should be large enough to reach convergence, which is typically only two or three. When multiple outer cycles, the number of inner cycles can be smaller as they do not necessarily help get overall convergence. Default value is 2.
Seasonal Smoother The length of the seasonal LOESS smoother.
Trend Smoother The length of the trend LOESS smoother.
Level Smoother The length of the level LOESS smoother.
Maximum Lag The maximum lag value. Use this parameter if the ACF is based on a range of lags. Default is 1.
Discrete Lags The comma-separated sequence of lag value. Use this parameter if the ACF should be based on discrete values. This sequence is empty by default.
Correlation Threshold The threshold used to determine the lag value with the highest correlation score. Default is 0.95.