Operator LPC

Primitive operator image not displayed. Problem loading file: ../../image/tk$com.teracloud.streams.timeseries/op$com.teracloud.streams.timeseries.modeling$LPC.svg

The Linear Predictive Coding (LPC) operator uses an autoregressive (AR) model to predict values in a time series. The LPC operator is a univariate operator and can ingest a univariate time series.

The AR model uses a linear predictive coding approach to forecast future time series values. The linear predictive coding approach is primarily used in audio and speech processing applications to encode and decode speech signals. This approach is also used in spectral estimation, smoothing, predicting stock values, and in gene sequence modeling.

The LPC operator supports a control port to which you can send control signals to control the behavior of the model. For example, you can make the model adaptive so that the predicted values reflect the trend of the data accurately. Using the control signals, the LPC operator can train the model to use a new input time series to predict values.

Behavior in a consistent region

  • The LPC operator is not supported in a consistent region. A warning occurs when you compile your streams processing application.
  • The operator cannot be the start of a consistent region. An error occurs when you compile your streams processing application.

Exceptions

The LPC operator does not throw any exceptions.

Examples

The following example shows how to forecast the next four values in a time series by using the LPC operator: 


composite Main {
  graph
    stream<int64 stampTime, list<float64> KPIs> KPIStream = FileSource()
    {
      param
        file : "KPIStream.csv";
        format:csv;
    }
    stream <timestamp predTimestamps, list<float64> 
    	predTimeSeries> predStream = LPC(KPIStream)
    {
      param
        inputTimeSeries: KPIStream.KPIs;
        initSamples: 144u;
        order: 8u;
        stepAhead: 4u;		  
        output predStream:
        predTimeSeries =forecastedTimeSeriesStep(),
        predTimes-tamps=forecastedTimestamps();
    }

    () as writer1 = File-Sink(predStream)
    {
      param
        file : "predstream.csv";
    }
}

Summary

Ports
This operator has 2 input ports and 2 output ports.
Windowing
This operator does not accept any windowing configurations.
Parameters
This operator supports 10 parameters.

Required: initSamples, inputTimeSeries

Optional: controlSignal, diagout, inputCoefficient, inputTimestamp, order, partitionBy, retrainingConfig, stepAhead

Metrics
This operator does not report any metrics.

Properties

Implementation
C++
Threading
Always - Operator always provides a single threaded execution context.

Input Ports

Ports (0)

This port consumes data for training and scoring against the model. The inputTimeSeries parameter specifies the name of the attribute on this port that contains the time series data. The accepted data types are float64 and list<float64>.

Properties

Ports (1)

This port accepts control signals to control the behaviour of the operator. The controlSignal parameter specifies the name of the attribute on this port that contains the control signal type. The inputCoefficient parameter specifies the attribute name on this port that contains the new input coefficents. The retrainingConfig parameter specifies the attribute name on this port that contains the configuration values used for retraining the model.

Properties

Output Ports

Assignments
This operator allows any SPL expression of the correct type to be assigned to output attributes.
Output Functions
LPC_COF
list<list<float64> > forecastedAllTimeSeriesSteps()

Returns a two-dimensional list of forecasted values where for each index in the input time series a list of forecasted values are returned (values from step 1 to n).

list<float64> forecastedTimeSeriesStep()

Returns the forecasted values at step n. The data at step n is a future time series value where n is the same as the stepAhead parameter value. If the input is a univariate time series, the output list contains a single forecasted value at step n.

<any T> list<T> forecastedTimestamps()

Returns a timestamp object that holds the value of forecasted timestamp values. If the input timestamp is irregular, the function returns a zero timestamp value and the operators logs a warning message in the log file. The supported types are uint64 and timestamp.

<any T> T AsIs (T v)

The default function for output attributes. By default, this function assigns the output attribute to the value of the input attribute with the same name.

Getcoeff
map<uint32,map<uint32,float64> > coefficients()

Returns the current coefficients of the LPC model.

<any T> T AsIs(T v)

The default function for output attributes. By default, this function assigns the output attribute to the value of the input attribute with the same name.

Ports (0)

This port submits a tuple that contains the next predicted value as determined by the LPC algorithm. This port submits a tuple each time a prediction is made. Custom output functions are used to specify the value of the output tuple attributes. The output tuple attributes whose assignments are not specified are assigned from input attributes.

Properties

Ports (1)

This port submits a tuple that contains information about the state of the LPC model. This port submits a tuple each time a Monitor signal is consumed on the input control port. The coefficients() output function is used to assign the value of the coefficients to an attribute. The expected type of the attribute is map<uint32,map<uint32,float64> >.

Properties

Parameters

This operator supports 10 parameters.

Required: initSamples, inputTimeSeries

Optional: controlSignal, diagout, inputCoefficient, inputTimestamp, order, partitionBy, retrainingConfig, stepAhead

controlSignal

Specifies the name of the attribute in the control port, which holds the control signal. The supported control signals are: TSSignal.Retrain, TSSignal.Monitor, TSSignal.Load, TSSignal.Suspend and TSSignal.Resume.

Properties

diagout

Get diagnostics to stdout

Properties

initSamples

Specifies the number of input time series values to be used to initialize the model.

Properties

inputCoefficient

Specifies the name of the attribute in the control port, which ingests the coefficients that are used for loading the model. If this parameter is not specified, by default, the inputCoefficient attribute is used. If the default attribute or the inputCoefficient parameter is not provided, the operator throws an exception. If the attribute or the parameter value does not contain valid coefficients, the load operation fails and the operator logs a warning message for each failed operation. The operator continues to predict values by using the older coefficients.

Properties

inputTimeSeries

Specifies the name of the attribute that contains the time series data in the input tuple. The supported types are float64 and list<float64>.

Properties

inputTimestamp

Specifies the name of the attribute in the input stream that contains the timestamp values. The supported data types are uint64 and timestamp. If the data type is uint64, the parameter value represents the number of nanoseconds since Unix epoch.

Properties

order

Specifies the order of the autoregressive model that the LPC operator uses for prediction. The default value is 24u.

Properties

partitionBy

Specifies the name of the attribute that contains the key values that are associated with the time series values in the input tuple. An input time series with the same partitionBy key is processed as a continuum. If the control port is specified, you must provide the key parameter for both the input ports. Otherwise, an exception is thrown.

Properties

retrainingConfig

Specifies the name of the attribute in the control port, which ingests the configurations that are used for retraining the model. If this parameter is not specified, by default, the retrainingConfig attribute is used. If the default attribute or the retrainingConfig parameter is not provided, the operator throws an exception. If the attribute or the parameter does not contain valid configurations, the retrain operation fails and the operator logs a warning message for each failed operation. The operator continues to predict values by using the older configuration.

Properties

stepAhead

Specifies the forecast horizon in the sample. The LPC operator produces a single-point forecast at the specified step. The default value is 1u.

Properties

Code Templates

LPC

stream<${schema}> ${outputStream} = LPC(${inputStream}) 
		{
		param
			inputTimeSeries:		${timeSeriesExpression};
			initSamples: ${initSamplesExpression};
		output
			${outputStream}: ${outputExpression};
		}

Libraries

LPC library module
Library Name: tsatapi
Library Path: ../../../impl/lib
Include Path: ../../../impl/include