Normalization by Min/Max

Table of contents

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

Introduction

Normalization by Min/Max rescales each spectrum to a 0-1 intensity range using its own minimum and maximum values.

How to use

To normalize your data by minimum & maximum:

  1. Upload data and select the spectra to be processed.
  2. Navigate to the sidebar and turn on the “Normalization” toggle.
  3. Select “Min max normalize” from the drop-down menu.
  4. Click “Process” to update the display.

Behavior

Normalizing by min and max is good for filtering data for machine learning and/or statistical analysis. This method shifts each spectrum downward so that the minimum is at 0, then rescales the data so that the maximum is at 1. Mathematically, the normalized intensity is derived by:

\[S(x)=\frac{I(x)-I_{\text{min}}}{I_{\text{max}}-I_{\text{min}}}\]

where $I_{\text{max}}$ is the peak intensity and $I_{\text{min}}$ is the minimum intensity.

Method

The implementation rescales each selected spectrum independently:

\[I_{\text{norm}}(x)=\frac{I(x)-\min(I)}{\max(I)-\min(I)}\]
Parameter Tunable or fixed Implementation
Method choice Tunable Selected as Min max normalize
Scaling bounds Fixed Per-spectrum minimum and maximum

References

  1. SpectraGuru implementation, self-implemented min-max normalization.

Back to Normalization