Open AccessCORRELATION AND REGRESSIONAL ANALYSIS OF THE DEPENDENCE OF ENTERPRISE RECEIPTS FROM THE FOREIGN ECONOMIC ACTIVITY FACTORS
Author(s) -
О. Ю. Малахова,
О. В. Кузнецова,
А. А. Кузнецов
Publication year - 2019
Publication title -
prikladnaâ matematika i voprosy upravleniâ
Language(s) - English
Resource type - Journals
eISSN - 2782-4500
pISSN - 2499-9873
DOI - 10.15593/2499-9873/2019.1.08
Subject(s) - revenue , regression analysis , econometrics , statistics , correlation coefficient , linear regression , total revenue , liberalization , economics , period (music) , business , actuarial science , mathematics , accounting , market economy , physics , acoustics
A correlation-regression analysis of the company's revenue dependence on the foreign economic activity factors was performed using the example of the Center for Appraisal and Analytical Technologies “Analytics.” The employees number (x1), rate $ (x2), rate € (x3), number of new open representative offices and branches for the period (x4) and service names in the company's nomenclature at the end of the period (x5) and number of suppliers at the end of the period (x6) revenue from the sale of medical equipment for the period (x7) and revenues from equipment maintenance and repair activities for the period (x8) and revenues from other activities for the period (x9), customs duties on equipment imports for the period (x10) and imports for the period (x11) and number of protectionism measures applied to the Russian Federation for the period (х12) and the number of measures of liberalization of foreign economic activity applied to the Russian Federation for the period (x13) was chosen as an external factor. The number of explanatory variables was reduced from 12 to 7 after the logical relationship between the factors analysis. Correlation analysis showed that many variables are intercorrelated. So, 8 regression models were constructed to avoid the negative effect of multicollinearity. These models do not contain strongly related factors. Statistical measures such as confidence intervals, the coefficient of determination, the Student’s index and the Fisher coefficient are given to assess the quality of the constructed models. It is shown that the constructed regression models are statistically significant. It’ll become possible to predict the results of the activity of the Center for Information Technologies "Analytics" in future periods if we used the regression models found. Also we could to minimize the risks due to the determination of the company's sources of income which determines the practical significance of the work.