Page 321 - IJB-9-6

International Journal of Bioprinting Rheology-informed machine learning model

Figure 3. Structure of artificial neural network-based models: (A) the parameter-dependent machine learning model, (B) the concentration-dependent
machine learning model, and (C) the rheology-informed hierarchical machine learning model. (D) Table of the machine learning model types and
corresponding model names.

model (rheology-informed hierarchical machine learning Additionally, the Keras module from TensorFlow was used
model [RIHML]) to predict the printing resolution using to develop the artificial neural network-based models,
regression as shown in Figure 3D. Two representative which specifically consisted of rectified linear unit (ReLU)
classical machine learning algorithms, RF and SVM, were function with the same hyperparameters, such as batch
implemented using the sci-kit learn package in Python. size and epoch. Furthermore, the train set, validation set,

Volume 9 Issue 6 (2023) 313 https://doi.org/10.36922/ijb.1280