Skip to main content

Numerics

Namespace: UTILSLIB  ·  Library: Utilities Library

#include <utils/numerics.h>

class UTILSLIB::Numerics

Static numerical helpers: integer GCD and log2, nchoose2, sparsity test, fixed-width histogram binning and time-series baseline rescaling shared across MATHLIB, DISP and the evoked / source-estimate pipelines.

Static scalar/per-vector numerical helpers (GCD, log2, histograms, baseline rescaling).

src/examples/ex_math/main.cpp
MatrixXd data(2, 6);
data << 1, 2, 3, 4, 6, 8, 2, 2, 2, 4, 4, 4;
RowVectorXf times(6);
times << -0.2f, -0.1f, 0.0f, 0.1f, 0.2f, 0.3f;
const MatrixXd percent = Numerics::rescale(data, times, std::make_pair(-0.2f, 0.0f), std::string("percent")); // (x - mean) / mean over baseline

Public Methods​

~Numerics()​

Destroys the Numerics object.

Static Methods​

gcd(iA, iB)​

Finds the Greatest Common Divisor (GCD) of two integer values.

Parameters:

  • iA : int First input integer.

  • iB : int Second input integer.

Returns:

  • int — The Greatest Common Divisor (GCD) of iA and iB.

issparse(v)​

Determines if a given data (stored as vector v) are representing a sparse matrix.

Parameters:

  • v : Eigen::VectorXd & Data to be tested.

Returns:

  • bool — true if sparse, false otherwise.

nchoose2(n)​

Calculates the combination of n over 2 (nchoosek(n,2)).

Parameters:

  • n : int The number of elements which should be combined with each other (n over 2).

Returns:

  • int — The number of combinations.

chi2Isf(p, dof)​

Value a chi^2 variable exceeds with probability p (scipy.stats.chi2.isf).

Parameters:

  • p : double Upper-tail probability.

  • dof : int Degrees of freedom.

Returns:

  • double — The quantile.

rescale(data, times, baseline, mode)​

Rescale (baseline correct) data.

Parameters:

  • data : const Eigen::MatrixXd & Data Matrix (m x n_time).

  • times : const Eigen::RowVectorXf & Time instants in seconds.

  • baseline : const QPair< float, float > & If baseline is (a, b) the interval is between "a (s)" and "b (s)". If a and b are equal, use interval between the beginning of the data and time 0.

  • mode : QString Baseline correction mode: "logratio" | "ratio" | "zscore" | "mean" | "percent".

Returns:

  • Eigen::MatrixXd — Rescaled data matrix.

rescale(data, times, baseline, mode)​

Rescale (baseline correct) data.

Parameters:

  • data : const Eigen::MatrixXd & Data Matrix (m x n_time).

  • times : const Eigen::RowVectorXf & Time instants in seconds.

  • baseline : const std::pair< float, float > & If baseline is (a, b) the interval is between "a (s)" and "b (s)". If a and b are equal, use interval between the beginning of the data and time 0.

  • mode : const std::string & Baseline correction mode: "logratio" | "ratio" | "zscore" | "mean" | "percent".

Returns:

  • Eigen::MatrixXd — Rescaled data matrix.

log2(d)​

Compute log base 2 of a given number.

Parameters:

  • d : const T Input value.

Returns:

  • double — log2 of d.

histcounts(matRawData, bMakeSymmetrical, iClassAmount, vecResultClassLimits, vecResultFrequency, dGlobalMin, dGlobalMax)​

Creates a class and frequency distribution from data matrix.

Parameters:

  • matRawData : const Eigen::Matrix< T, Eigen::Dynamic, Eigen::Dynamic > & Raw data matrix that needs to be analyzed.

  • bMakeSymmetrical : bool User input to turn the x-axis symmetric.

  • iClassAmount : int User input to determine the amount of classes in the histogram.

  • vecResultClassLimits : Eigen::VectorXd & The upper limit of each individual class.

  • vecResultFrequency : Eigen::VectorXi & The amount of data that fits in the appropriate class ranges.

  • dGlobalMin : double User input to determine the minimum value allowed in the histogram.

  • dGlobalMax : double User input to determine the maximum value allowed in the histogram.

histcounts(matRawData, bMakeSymmetrical, iClassAmount, vecResultClassLimits, vecResultFrequency, dGlobalMin, dGlobalMax)​

histcounts(matRawData, bMakeSymmetrical, iClassAmount, vecResultClassLimits, vecResultFrequency, dGlobalMin, dGlobalMax)​

Example​

Used in examples​

ex_histogram​

Example application of reading raw data and presenting the result in histogram form.

src/examples/ex_histogram/main.cpp, line 204
dataSine = sineWaveGenerator(1.0e-6, (1.0 / 1.0e6), 0.0, 1.0); //creates synthetic data using sineWaveGenerator function
qDebug() << "sourceEstimateData.rows = " << sourceEstimateData.rows();
qDebug() << "sourceEstaimateData.cols = " << sourceEstimateData.cols();
Numerics::histcounts(sourceEstimateData, bMakeSymmetrical, classAmount, resultClassLimit, resultFrequency, inputGlobalMin, inputGlobalMax); //user input to normalize and sort the data matrix
std::cout << "resultClassLimits = " << resultClassLimit << std::endl;
std::cout << "resultFrequency = " << resultFrequency << std::endl;
qDebug() << "HistCounts timer:" << myTimerHistCounts.elapsed();

Authors of this file​