Active learning for prediction of tensile properties for material extrusion additive manufacturing

Training dataset

The training dataset, which was taken from Pourali and Peterson as well as experiments that led to the determination of print parameters to use in that study, was first investigated for correlations between print parameters (inputs) and mechanical properties (outputs) and for correlations between different mechanical properties³⁶. A heatmap, shown in Fig. 3, was generated using the algorithm available in the seaborn library.

T_ext is positively correlated with y2 and x2,…