reference
labkit.biosignal.compareGroups
Summarize groups and compute pairwise Welch comparisons.
Syntax
result = labkit.biosignal.compareGroups(values, groups)Description
Calculates descriptive statistics for each group and a two-sided Welch t-test for every pair of groups. Welch's test does not assume equal variances. Group order follows the first occurrence of each label in groups.
Rows with a nonfinite measurement or a blank group label are discarded before calculation. A pairwise row is still returned when either group has fewer than two observations or the estimated standard error is zero; its test statistic, degrees of freedom, and p-value are NaN.
Inputs
values- Numeric vector of measurements.
groups- Vector of labels accepted by string, such as a string array, cell array of character vectors, or categorical array. It must contain one label for each element of values.
Outputs
result- Structure with summary and pairwise tables.
Output Fields
summary- One row per group. Columns are Group, N, Mean, Std, Median, Min, and Max.
pairwise- One row per unique group pair. Columns are GroupA, GroupB, MeanDifference, T, DF, and P. MeanDifference is mean(GroupA) minus mean(GroupB); P is the two-sided Welch-test p-value.
Failure Behavior
Groups with fewer than two finite observations, or with zero estimated standard error, retain a pairwise row with NaN test statistics. values and groups must contain the same number of elements; incompatible input shapes or values that cannot be converted to numeric/text raise the originating MATLAB error.
Example
values = [4.8 5.1 5.0 6.2 6.0 6.4];
groups = ["control" "control" "control" "treated" "treated" "treated"];
result = labkit.biosignal.compareGroups(values, groups);Related APIs
labkit.biosignal.measureSegments— Measure generic template-residual segment quality.labkit.biosignal.buildTemplate— Build a representative segment template.labkit.biosignal.cropSignal— Return a signal clipped to a time range in seconds.labkit.biosignal.defaultEcgPeakOptions— Return documented defaults for ECG/QRS peak detection.labkit.biosignal.detectEcgPeaks— Detect ECG/QRS peaks as event anchors.labkit.biosignal.filterSignal— Apply a zero-phase FFT-domain filter to a biosignal.labkit.biosignal.getChannel— Return one signal from a recording by display name or index.labkit.biosignal.listChannels— Return display names for all channels in a recording.
Source
This page is generated from the MATLAB help text in +labkit/+biosignal/compareGroups.m.