a comparative study of conventional machine leaning and neural networks for predicting mechanical properties of bone scaffolds

As calcium phosphate/polymer composite scaffolds are prevalently used in bone tissue engineering research, an accurate prediction of their elastic moduli based on the material selection, contributes toward their clinical success rate. in this study, the conventional machine learning regression tools such as multiple linear (lr), ridge (rg), random forest (rf), decision tree regression (dtr), xgboost (xgb), as well as an artificial neural network (ann) were used to predict the elastic moduli of 3d composite scaffolds composed of calcium phosphates and polymers (such as alginate, chitosan, polycaprolactone, etc.) for bone tissue engineering applications. in conclusion, xgb and ann have been found to be the best models that could predict the elastic modulus with the highest r2 scores.