Efficient and accurate determination of the degree of substitution of cellulose acetate using ATR-FTIR spectrosopy and machine learning
Frank Rhein, Timo Sehn, Michael A.R. Meier
This repository contains the data evaluation reported in "Efficient and accurate determination of the degree of substitution of cellulose acetate using ATR-FTIR spectrosopy and machine learning" (https://doi.org/10.1038/s41598-025-86378-0).
Data used in this repository is published under the CC BY-NC 4.0 license here: https://publikationen.bibliothek.kit.edu/1000172511 (DOI: 10.35097/tvwlylbMvDXhEcRt) and simple downloaded and extracted to the data/ folder in this repository.
All code is written in Python. The reported study is contained in ds_ir_ml.py and is divided into sections according to the publication. Each section can be run individually by setting the corresponding boolean values in lines 25-30.
More detailed calculations and the config class are contained in mod/util_functions.py. In config, the default evaluation parameters can be set. Most importantly N_REPS, defining the number of repetitions. The study uses 1000, however for quick access choose a smaller value. Note that exact results are stochastic in nature (random test/train split).
A custom plotter function is used and provided in mod/custom_plotter.py.