ANN model PHB production acetic acid supplementation

This is a Digital workflow for modelling PHB production with acetic acid, and time of addition as predictors. The scripts used are for optimizing the ANN architecture/data split. Analyzing the optimized network for fitness, plotting response surface, LIME and SHAP analysis

本数字工作流以乙酸与物料添加时间作为预测变量，用于构建聚羟基丁酸酯（Polyhydroxybutyrate, PHB）的生产模型。所使用的脚本用于优化人工神经网络（Artificial Neural Network, ANN）的架构及数据划分方案，并将针对优化后的网络开展拟合度分析、响应面绘制、局部可解释模型无关解释（Local Interpretable Model-agnostic Explanations, LIME）以及SHAP（SHapley Additive exPlanations）分析。