prediction of compressive strength of geopolymer fiber reinforced concrete using machine learning

Geopolymers represent a cutting-edge class of inorganic materials that provide a sustainable substitute for conventional cement and concrete. through meticulous combinations and ratios of elements like fly ash (fa), silica fume, ground granulated blast slag (ggbs), alkaline solutions, aggregates, superplasticizers, and fibers, geopolymer concrete mixes are generated as part of the experimental program. the investigation concentrates on predicting the 28-day compressive strength, a pivotal parameter in assessing concrete performance. the dataset comprises 96 data points, and two advanced techniques, namely support vector regression (svr) and artificial neural networks (ann), are harnessed for this research. the ann demonstrates an  value of 0.992 on the training dataset, indicating its capacity to elucidate around 99.2% of the variability. on the other hand, svr boasts an  value of 0.995, signifying an ability to account for about 99.5% of the variance. when applied to the testing data, the ann achieves an  of 0.96, while svr attains an  of 0.99. this study suggests that svr exhibits slightly superior performance in elucidating variance within the testing dataset.