期刊名称：Eastern-European Journal of Enterprise Technologies
印刷版ISSN：1729-3774
电子版ISSN：1729-4061
出版年度：2017
卷号：2
期号：3
页码：51-59
DOI：10.15587/1729-4061.2017.97969
语种：English
出版社：PC Technology Center
摘要：In the practical cultivation of cultures, including grapes, prediction is an inherent attribute, since, under weather actions, it is possible to immediately take measures before one or another degree of the vegetation of culture is affected. In order to develop mathematical software of an information system for determining the probability of reduction in the yield of grapes, we observed the annual variation in temperatures. Moreover, to manage this process, its comparison was conducted with the development of grapevine by the phases of its development for 16 years. By employing the method of binary logistic regression, we revealed three most significant indicators (radiation background, the sum of efficient temperatures during flowering, annual total precipitation in the previous year), which were included into the mathematical model developed. The results obtained make it possible to estimate the risk of reduction in the harvest of grapes, which is grown under conditions of the Northeastern forest-steppe region of Ukraine (Kharkiv region). The developed model as a whole and its separate coefficients are statistically significant. It is also established that all the predictors, in accordance with the chi-squared test, impact the prediction of reduction in the yield of grape. The obtained results might be used when making a decision about the need of change in the agrotechnical methods for the purpose of improving productivity by changing the course of specific phases in the vegetation of grape.
关键词：harvest prediction;logistic regression;information system;control over yield;a grape culture
