Amirkabir University of Technology Amirkabir Journal of Civil Engineering 2588-297X 56 12 2025 02 19 Development of innovative intelligent model to estimate the strength properties of hemp bio composite mixture using the combination of water cycle algorithm and MARS method Development of innovative intelligent model to estimate the strength properties of hemp bio composite mixture using the combination of water cycle algorithm and MARS method 1557 1582 5586 10.22060/ceej.2024.22940.8080 FA Davood Sedaghat Shayegan Department of civil engineering, Roudehen Branch, Islamic Azad University, Tehran, Iran Elahe Fazlikhani Islamic, Azad University, Roudhen, branch, Roudhen, Iran َAli Asghar Amirkardoust Department of civil engineering, islamic azad university, roudehen branch, roudehen, iran Journal Article 2024 01 21 Considering the characteristics of ordinary concrete, which is a composite material with characteristics such as resistance and low tensile strain range. Basically, the two main weaknesses of concrete, which include the brittle behavior of concrete and the weakness in its elasticity, have made the use of concrete structures made of ordinary concrete face major problems. In general, by adding fibers to the concrete mixture, it is possible to improve the mechanical properties of concrete. In this research, an innovative approach using the water cycle algorithm was used to combine with the adaptive multivariate regression spline method (MARS) to model and predict the compressive strength and tensile strength of concrete containing hemp biocomposites. For the development of each of the proposed models, 153 mixing designs along with their compressive strength results were collected from authoritative articles. After analyzing and evaluating the influencing parameters by the Mallow coefficient, the input parameters to the smart models include the ratio of water to cement base materials, the ratio of hemp seeds to cement base materials, the weight percentage of hemp seeds, cement base materials, seeds hemp, density of cement base material, density of dry material and strength of cement base material were selected. The results showed that the correlation coefficient for the compressive strength model for MARS optimized with the algorithm and Mars is 0.991 and 0.971, respectively, and the tensile strength is 0.928 and 0.911, respectively, in the training and testing phase. Investigations show that the proposed MARS model optimized with a meta-heuristic algorithm has good performance and high accuracy in estimating the compressive strength and tensile strength of concrete containing hemp biocomposite. The external validation results also show that the proposed approach can be introduced as predictive models and the correlation between predicted values ​​and experimental values ​​cannot be random. Considering the characteristics of ordinary concrete, which is a composite material with characteristics such as resistance and low tensile strain range. Basically, the two main weaknesses of concrete, which include the brittle behavior of concrete and the weakness in its elasticity, have made the use of concrete structures made of ordinary concrete face major problems. In general, by adding fibers to the concrete mixture, it is possible to improve the mechanical properties of concrete. In this research, an innovative approach using the water cycle algorithm was used to combine with the adaptive multivariate regression spline method (MARS) to model and predict the compressive strength and tensile strength of concrete containing hemp biocomposites. For the development of each of the proposed models, 153 mixing designs along with their compressive strength results were collected from authoritative articles. After analyzing and evaluating the influencing parameters by the Mallow coefficient, the input parameters to the smart models include the ratio of water to cement base materials, the ratio of hemp seeds to cement base materials, the weight percentage of hemp seeds, cement base materials, seeds hemp, density of cement base material, density of dry material and strength of cement base material were selected. The results showed that the correlation coefficient for the compressive strength model for MARS optimized with the algorithm and Mars is 0.991 and 0.971, respectively, and the tensile strength is 0.928 and 0.911, respectively, in the training and testing phase. Investigations show that the proposed MARS model optimized with a meta-heuristic algorithm has good performance and high accuracy in estimating the compressive strength and tensile strength of concrete containing hemp biocomposite. The external validation results also show that the proposed approach can be introduced as predictive models and the correlation between predicted values ​​and experimental values ​​cannot be random. https://ceej.aut.ac.ir/article_5586_8153349b1eb6f6c01622d4f395993bfe.pdf