fig4

From: Defect engineering and machine learning enhance n-butanol sensing of γ-Fe₂O₃

Defect engineering and machine learning enhance <i>n</i>-butanol sensing of γ-Fe<sub>2</sub>O<sub>3</sub>

Figure 4. (A) Corrected concentration using extra trees (ETs); (B) Empirical cumulative distribution function of calibration errors; (C) Target Plot of ML Models; (D) Confusion matrix heatmaps for gas classification using four sensors, corresponding to the CatBoost algorithm; (E) Accuracy plot of the algorithms; (F) Feature analysis plot for the four sensors. All subfigures are original. Panels (A), (B), (C), (E), and (F) were plotted using Origin 2024. Panel (D) was plotted using Python/matplotlib. NN: Neural network; DT: decision tree; RF: random forest; ET: extra tree; LR: linear regression; RMSE: root mean square error; ACE: acetone; ACN: acetonitrile; DMF: N, N-dimethylformamide; IPA: isopropyl alcohol; NBA: n-butanol; TEA: triethylamine; TMA: trimethylamine; BP: Backpropagation; SVM: Support Vector Machine; ML: machine learning.