CROWDFUNDING SYSTEMS: evaluation and government regulation in the conditions of reengineering

A model is a simplified representation of a real object, process, or phenomenon. We used mathematical models, i.e. systems that are formalized. So, it is a set of interrelated mathematical and formal-logical expressions that usually reflect real processes and phenomena ( Ȼɥɚɝɭɧ et al., 2011). Besides, a dynamics chart, a trend line, and so on were created for each of the variables. Numerical models allow to get the solution of analytical models and are applied to specific output data and have an additional error. These models are universal and convenient for solving complex problems, but they are not visual and time-consuming while analyzing and establishing relationships between model parameters. Using regression analysis, you can continue the trend line to predict future values ( Ȼɥɚɝɭɧ et al., 2011; Ɏɟɳɭɪ et al., 2003). Simple linear regression establishes a linear relationship between two variables. One of them (y) is considered as a result variable and is a function of the regressor (x). Value ɭ ={ ɭ 1, ɭ 2…, ɭ n} consists of 2 components: 1) the non-random component Į + ȕ x, where ɯ ={ ɯ 1, ɯ 2,..., ɯ n}, Į and ȕ are the parameters of the equation; 2) the random component e={e1,e2,..,en} ( Ȼɥɚɝɭɧ et al., 2011). Errors are usually the result of a multidimensional normal distribution. The constant a defines the intersection point of the regression line with the ordinate axis and is the average value at the point x 0 = 0. The coefficient b characterizes the slope of the straight line to the axis of the abscissus ( Ȼɥɚɝɭɧ et al., 2011; Ɏɟɳɭɪ et al., 2003). The quality of each regression model was checked by the Fisher criterion and the coefficient of determination. For such a dependent variable as the number of crowdfunding platforms, all the proposed regression models confirm their adequacy and their coefficient of determination is greater than 0.83 (Table 2.3). It was found that when analyzing the number of funds raised through crowdfunding and the number of funds raised through a crowdfunding platform, only half of the models are adequate. Based on the results of the study of basic and modified models of linear, logarithmic, hyperbolic and degree types, and based on calculations of model parameters based on actual data, it was found that the most adequate models for describing the dependence of the number of crowdfunding platforms are a degree model, and for describing the dependence of the amount of funds with the time factor — a logarithmic (exponential) model. 56