org.hawkular.datamining.forecast.models

Interface ModelOptimizer

All Known Implementing Classes:

AbstractModelOptimizer, DoubleExponentialSmoothing.DoubleExOptimizer, SimpleExponentialSmoothing.SimpleExOptimizer, TripleExponentialSmoothing.TripleExOptimizer

public interface ModelOptimizer

Finds best model for given data set. Time series model in general contains parameters which optimal values can be chosen to get model which best fits the data.

Author:

Pavol Loffay

Method Summary

Modifier and Type	Method and Description
TimeSeriesModel	minimizedMSE(List<DataPoint> trainData)
double[]	result()

Method Detail

minimizedMSE

TimeSeriesModel minimizedMSE(List<DataPoint> trainData)

Parameters:

trainData - training data set

Returns:

best model for given data set. Model is already learned up to last timestamp in trainData

result

double[] result()

Returns:

optimized parameters of the model. Length of array depends on a model being optimized.