The use of the inverse transformation method for time series analysis

In modern conditions of technology development, signs of systemacity are manifested to one degree or another in all areas, so the use of system analysis is an urgent task. In this case, the main factors in this situation are data processing and prediction of the state of a system. Mathematical modeling is used as a prediction method for a given subject area. A mathematical model is a universal tool for describing complex systems representing the approximate description of the class of phenomena of the external world expressed by mathematical concepts and language. The mathematical model can be represented as a set of systematic components and a random component. In this article, the object of prediction is the irregular random component of a model, which reflects the impact of numerous random factors. The origin, nature and laws of variation of the random variable are known, therefore, to simulate its behavior or predict its future value, one needs high degree of certainty to establish the form of continuous distribution function of the random variable. The empirical distribution function is calculated using the sample of random variable values. This empirical function is close to the values of the desired unknown function of distribution. The resulting empirical function is discrete, therefore it is necessary to apply piecewise linear interpolation to obtain a continuous distribution function. 

The predicted random component of time series has been included in the initial regression model. In order to compare augmented and initial regression models, several values were excluded from the time series and new prediction was built. The value of the average approximation error for assessing the quality of the model is calculated. The augmented regression model proved to be more effective than the original one.