Normalization by Mean

Table of contents

  1. Introduction
  2. How to use
  3. Behavior
  4. Method
  5. References

Introduction

Normalization by Mean rescales each spectrum by its own average intensity. It is useful when spectra should be compared relative to their typical signal level rather than their absolute intensity.

How to use

  1. Upload spectra and open Processing Page.
  2. Turn on Normalization in the sidebar.
  3. Choose Normalize by mean from Select normalization function.
  4. Apply preprocessing and review the normalized spectra.

Behavior

SpectraGuru calculates the arithmetic mean of all intensity values in each spectrum column, then divides that column by its own mean. The wavenumber axis is unchanged. There are no additional parameters for this normalization method. A spectrum needs a nonzero mean for the division to produce finite values.

Method

For a spectrum with intensities $I_1,\ldots,I_N$, the normalized intensity at point $i$ is

\[I_i^{\mathrm{norm}}=\frac{I_i}{\bar I},\qquad \bar I=\frac{1}{N}\sum_{j=1}^{N}I_j.\]
Parameter Tunable or fixed Implementation
Normalization function Tunable Select Normalize by mean
Mean intensity Fixed Arithmetic mean calculated separately for each spectrum
Wavenumber axis Fixed Unchanged

References

  1. NumPy Developers. numpy.mean.