Построение регрессионных моделей кривых обучения для промышленных предприятий

The problem of constructing regression models of learning curves is considered based on the statistical data of the Salyut JSC enterprise on the production of new products.Using the nonlinear method of least squares, the parameters of five models of learning curves of a given type are defined: Wright's, Stanford-B, Deyong's, logistic and exponential.As a result of the statistical characteristics analysis of the constructed models, the regression dependence is chosen that best approximates the data.