Examples

These examples demonstrate how to use the VAR2 operator.

The following example demonstrates multivariate prediction of IBM stock prices and trading volumes:


use com.teracloud.streams.timeseries.modeling::VAR2;
composite Main {
  graph
    stream<list<float64> inp> StockInp = FileSource() {
      param
        file: "stockInp.csv";  // read stock data
        format: csv;
        initDelay: 1.0f;	
    }
    
    stream<list<float64> predictions> varOut = VAR2(stockIp) {
      param
        inputTimeSeries: inp;  // time series to be predicted
        initSamples: 12u;      // use 12 early data to train the model
        updateTime: 6u;        // update the model every 6 samples
      output
        varOut: predictions = predictedTimeSeries(); // get prediction value for all time series component
    }
    
    () as FlSnk = FileSink(varOut) {
      param
        file:"Decomposition.dat";
    }
}

The following graphs illustrate the actual and predicted stock prices and trade volumes for a sample time series: