Skip to contents

Dimensionality Reduction for Biomechanics

Source: R/ops-dimred.R
extractWaveformFeatures.Rd

Functions for dimensionality reduction and visualization of high-dimensional biomechanical waveform data including PCA, UMAP, and t-SNE. Extract waveform features for dimensionality reduction

Usage

extractWaveformFeatures(x, features = c("statistical", "shape"), n_points = 50)

Arguments

x

A PhysioExperiment object or matrix (time x observations).

features

Character vector of features to extract: "raw" (flattened waveform), "statistical" (summary stats), "frequency" (spectral features), "shape" (curve characteristics).

n_points

For "raw", number of points to resample to.

Value

A matrix (observations x features) suitable for PCA/UMAP.

Details

Extracts summary features from waveforms for use with standard dimensionality reduction methods.

Feature types:

  • raw: The raw waveform resampled to n_points

  • statistical: Mean, SD, min, max, range, skewness, kurtosis

  • frequency: Dominant frequency, spectral centroid, bandwidth

  • shape: Peaks, zero crossings, area under curve

References

van der Maaten L, Hinton G (2008). "Visualizing Data using t-SNE." Journal of Machine Learning Research, 9, 2579-2605.

See also

waveformPCA(), waveformUMAP(), waveformTSNE()

Examples

# Extract features from gait data
set.seed(123)
data <- matrix(rnorm(1000), nrow = 100, ncol = 10)
features <- extractWaveformFeatures(data, features = c("statistical", "shape"))