Fig. 10: Classifier design and accuracy evaluation results for 3 different classification tasks. | Microsystems & Nanoengineering

Fig. 10: Classifier design and accuracy evaluation results for 3 different classification tasks.

From: Machine learning metrology of cell confinement in melt electrowritten three-dimensional biomaterial substrates

Fig. 10

Support vector machines are used for all the machine learning tasks. a Multi-class classification task taking into consideration all classes. Scatter plot for the combination of metrics. Blue, red, pink, and green points represent processed cells from Class A, B, C, and D, respectively. 88 cells are processed whereby the computed are normalized and then plotted to assess the classifier design results. b Confusion matrix with average classification accuracy outlined with yellow color (64.4%). c Multi-class classification task taking into consideration only Class A, B, and C. d Confusion matrix with average classification accuracy outlined with yellow color (90.6%). e Binary classification task by combining Classes A, B, and C in one class designated as Class ABC against Class D. f Confusion matrix with average classification accuracy outlined with yellow color (93%)

Back to article page