Utilization of Gene Expression Programming for Modeling of Mechanical Performance of Titanium/Carbonated Hydroxyapatite Nanobiocomposites: the Combination of Artificial Intelligence and Material Science

Titanium carbonated hydroxyapatite (ti/cha) nanobiocomposites have extensive biological applications due to the excellent biocompatibility and similar characteristics to the human bone. ti/cha nanobiocomposite has good biological properties but it suffer from diverse characteristics especially in hardness, young's modulus, apparent porosity and relative density. this investigation is an attempt to propose the predictive models using gene expression programming (gep) to estimate these characteristics. in this regards, gep is used to model and compare the effect of practical variables including pressure, ti/cha contents and sintering temperature on their monitored properties. to achieve this goal, 90 different experiments were considered to create the gep models. selected data set were divided randomly into 63 training sets and 27 testing sets. finally, five of the best models were reported for each different output. sensitivity analyses were done to determine and rank the practical parameters on each of the investigated properties and revealed that wt.% ti, wt.% cha, compaction pressure (mpa) and temperature (ᵒc), respectively are the most effective parameters on hardness, young's modulus, shear modulus, apparent porosity and relative density. by comparing the results, a very good agreement was observed between the experimental data and the results obtained from gep model.