ML model prediction to optimize composition of cellulose based film

Published: 5 March 2025| Version 1 | DOI: 10.17632/53yfxbcdbj.1
Contributor:
BISWANATH MAHANTY

Description

This data set contains properties of cellulose-based film prepared with different proportions of PEG, malic acid, and hexdecanoate. Four attributes of the film are modeled with Fitlm (ANN) and ensemble modeling tools (RF) while optimizing hyperparameters. The composite desirability-based optimization workflow is presented.

Files

Steps to reproduce

The entire work flow, developed from dataset mydata.mat, has been presented in the modeling_exercise.m script. The script calls for all functions (script directory) to develop the ML models: individual optimization, multiobjective optimization. Additionally, LIME, SHAP explanation, object creation, and plotting are included.

Institutions

Karunya University

Categories

Artificial Neural Network, Machine Learning, Multi-Objective Optimization, Ensemble, Food Packaging Material, Random Decision Forest, Shapley Value, Interpretable Machine Learning

Licence