Method for Statistical Evaluation of Nonlinear Model Parameters in Statistical Learning Algorithms
摘要
This paper proposes a method for determining the parameters of nonlinear models in statistical learning algorithms. The proposed solution allows to determine the parameters of a mathematical model with significant nonlinear components from statistical discrete data. Unlike classical approaches, the method allows to obtain the final analytical solution. This is achieved by using the method of differential transformations. As a result, we will have a statistically trained model with high predictive properties.