Estimation of the Mean Grain Size of Mechanically Induced Hydroxyapatite Based Bioceramics via Artificial Neural Network

This study focuses on the estimation of the mean grain size of mechanically induced Hydroxyapatite (HA) through the artificial neural network (ANN) model. The mean grain size of HA and HA based nanocomposites at different milling parameters was obtained from previous studies. The data were trained and tested by the neural network modeling. Accordingly, all data (55 sets) were based on mechanical alloying and were randomly divided into 40 training sets and 15 testing sets. The data used in the multilayer feed forward neural networks models and input variables of models were arranged in a format of 13 input parameters. The results indicated a very good agreement between the experimental data and the predicted ones. The R2 value of the trained and tested data suggested by the model confirmed this situation. Given the broad range of the used parameters, it was found that our analysis and model were fully functional to accurately estimate the optimal conditions for experiments. This shows the potential application of these calculations and analyses in a wide range of numerical studies.