Abstract:
Mathematical models are important not only to provide a better understanding for the underlying dynamics of huge systems, whose components are highly correlated, but also enables us to interpret their future behavior and compare with other possible situations. In this respect, there is a crucial demand for efficient new approaches to infer such complex phenomena which may appear in real-world problems of the fields like system biology, medicine, engineering, finance, education, environment and so on. In such systems, the components appear in a high dimensional network structure where they are interconnected with each other and under the effect of some unknown external parameters needed to be identified. In this study, we implement conic multivariate adaptive regression spline (CMARS) technique on a real-world data from system biology for the inference of such complex regulatory networks. The performance of the model is investigated in comparison with the results obtained from multivariate adaptive regression spline (MARS) modeling approach.