comparing time series data

I'd like to compare two time series data sets. These data sets measure behavior in response to a stimulation in test and control animals. I'd like to test wether my test line is statistically different from a control line at which time point. Do you have any suggestions of the best way of doing this?

[Yuchen Chen]

simulinktest.constraints.TimeseriesSignalComparator

TimeseriesSignalComparator - Compares Timeseries signals or arrays of

Timeseries signals

By default the TimeseriesSignalComparator compares two Timeseries

signals or Timeseries signal arrays for numerical equality. The signal

numerical equality is qualified by comparing 'Time' and 'Data' fields of

the objects (or arrays of objects). In order to compare all displayed

fields, set 'ComparingAllFields' option to true as shown in 'Example 2:

Comparing all displayed fields'.

For regular comparison of two timeseries objects, use IsEqualTo without

any comparator specification. Refer to help on

matlab.unittest.constraints.IsEqualTo for more information on comparing two

objects in general.

TimeseriesSignalComparator performs the following preliminary checks

prior to signal value comparison:

* Both actual and expected values are Timeseries objects

* Both actual and expected have the same number of Timeseries objects

* Start and stop times of actual and expected Timeseries signals are

the same

* The actual signal's time vector contains at least all time points

contained in the expected signal's time vector

Failing to satisfy any of these requirements will prevent signal value

comparison from happening and will result in an early termination with

the appropriate qualification failure message.

Signals are assumed continuous by default and thus will be compared at

all time points in the expected time vector. If signals are

discontinuous and comparisons around zero crossings need to be ignored,

set 'IgnoringSignalValuesAtZeroCrossings' option to true as shown in

'Example 3: Ignoring comparison around signal discontinuities'.

See 'IgnoreSignalValuesAtZeroCrossings' property for information on how

signal discontinuities are determined.

To compare signal values with a tolerance, specify the kind of

tolerance desired as shown in 'Example 4: Comparison using tolerance'.

Refer to documentation on MATLABUnit Qualification API to obtain a list

of supported tolerance types.

TimeseriesSignalComparator properties:

CompareAllFields - Boolean indicating whether or not to compare all displayed fields to prove equality

IgnoreSignalValuesAtZeroCrossings - Boolean indicating whether or not to ignore comparison around points of signal discontinuity

TimeseriesSignalComparator methods:

TimeseriesSignalComparator - Class constructor

Usage:

testcase.verifyThat(actual, IsEqualTo(expected, 'Using', TimeseriesSignalComparator))

Examples:

To use any example, first execute the following statements

import simulinktest.constraints.TimeseriesSignalComparator

import matlab.unittest.constraints.AbsoluteTolerance

import matlab.unittest.constraints.IsEqualTo

%%% 1. Simple equality check %%%

testcase.verifyThat(timeseries(1:10, 1:10), IsEqualTo(timeseries(1:10, 1:10), 'Using', TimeseriesSignalComparator));

%%% 2. Comparing all displayed fields (Failing Example)%%%

exp = timeseries(1:10,'Name', 'expected');

act = timeseries(1:10,'Name', 'actual');

testcase.verifyThat(act, IsEqualTo(exp, 'Using', TimeseriesSignalComparator('ComparingAllFields', true));

%%% 3. Ignoring comparison around signal discontinuities %%%

testcase.verifyThat(timeseries(1), IsEqualTo(timeseries(1), 'Using', TimeseriesSignalComparator('IgnoringSignalValuesAtZeroCrossings', true)))

%%% 4. Comparison using tolerance %%%

testcase.verifyThat(timeseries(1), IsEqualTo(timeseries(1.01), 'Using', TimeseriesSignalComparator, 'Within', AbsoluteTolerance(0.1)));

See also

simulinktest.constraints.SimulinkOutputComparator matlab.unittest.constraints.IsEqualTo matlab.unittest.constraints.Tolerance

NOTE

If ignoring time points of discontinuity produces an empty resultant time vector, comparison will result in a qualification failure.

Class Details

Superclasses matlab.unittest.internal.constraints.ContainerComparator, simulinktest.internal.mixin.IgnoreSignalValuesAtZCMixin, simulinktest.internal.mixin.SignalValueAlignmentMixin, simulinktest.internal.mixin.TimeseriesMinerMixin, simulinktest.internal.mixin.CompareAllFieldsMixin, simulinktest.internal.mixin.HasTimeTolerance, simulinktest.internal.mixin.WithinRecorderMixin

Sealed false

Construct on load false

Constructor Summary

TimeseriesSignalComparator

Property Summary

CompareAllFields Indicates whether to compare all fields of a

IgnoreSignalValuesAtZeroCrossings Indicates whether to ignore

IsStateful

Recursive Boolean indicating whether the instance operates recursively

SignalAlignment Indicates how to align two signals that need to

TimeTolerance Adds a TimeTolerance property that can be used by

Method Summary

protected Static extractSignalValues Extracts signal values from a Timeseries formatted Simulink

protected Sealed getComparatorsForElements

protected isStateful

Answer 1

[Chad Greene]

[R,P] = corrcoef(testData,controlData); will return a correlation between your two data sets and the probability that such a correlation coefficient could be obtained by chance.