org.wso2.siddhi.extension.timeseries.linreg

Class LengthTimeMultipleLinearRegression

java.lang.Object

org.wso2.siddhi.extension.timeseries.linreg.LengthTimeRegressionCalculator

org.wso2.siddhi.extension.timeseries.linreg.LengthTimeMultipleLinearRegression

public class LengthTimeMultipleLinearRegression
extends LengthTimeRegressionCalculator

Field Summary

Fields inherited from class org.wso2.siddhi.extension.timeseries.linreg.LengthTimeRegressionCalculator

batchSize, calcInterval, confidenceInterval, currentEventTime, duration, eventCount, incCounter, xParameterCount

Constructor Summary

Constructors

Constructor and Description
LengthTimeMultipleLinearRegression(int paramCount, long timeWindow, int lengthWindow, int calcInt, double ci) This method calls the constructor of super class

Method Summary

Modifier and Type	Method and Description
protected void	addEvent(Object[] data, long expiryTime) Method to accumulate new events: dependent (Y) and independent (X) variable values
protected Object[]	processData() Method which performs multiple linear regression
protected void	removeExpiredEvents() Method to remove events which have expired based on time and/or length

Methods inherited from class org.wso2.siddhi.extension.timeseries.linreg.LengthTimeRegressionCalculator

calculateLinearRegression

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

LengthTimeMultipleLinearRegression

public LengthTimeMultipleLinearRegression(int paramCount,
                                          long timeWindow,
                                          int lengthWindow,
                                          int calcInt,
                                          double ci)

This method calls the constructor of super class

Parameters:

paramCount - Number of X variables + 1 (Y variable)

timeWindow - Time window which is used to constrain number of events

lengthWindow - Length window which is used to constrain number of events

calcInt - Frequency of regression calculation

ci - Confidence interval

Method Detail

addEvent

protected void addEvent(Object[] data,
                        long expiryTime)

Method to accumulate new events: dependent (Y) and independent (X) variable values

Specified by:

addEvent in class LengthTimeRegressionCalculator

Parameters:

data - Array which holds Y and X values

expiryTime - Time at which each event should expire (based on time window)

removeExpiredEvents

protected void removeExpiredEvents()

Method to remove events which have expired based on time and/or length

Specified by:

removeExpiredEvents in class LengthTimeRegressionCalculator

processData

protected Object[] processData()

Method which performs multiple linear regression

Specified by:

processData in class LengthTimeRegressionCalculator

Returns:

Array containing standard error and beta values