I. M. Grod, L. O. Shevchik, N. Ya. Kravets
Volodymyr Hnatiuk Ternopil National Pedagogical University, Ukraine, HorbachevskyTernopil State Medical University, Ukraine
THE USE OF COMPUTER MODELING FOR THE PREDICTION OF THE DYNAMICS OF INSECT-ANTHOPHILIAN POPULATIONS
The paper describes the algorithm for constructing ARIMA based on the Box-Jenkins prediction method. The analysis of long-term observation series is one of the main tasks of environmental monitoring. In environmental monitoring programs, an important place is given to the development of methods for modeling population dynamics, as well as exploring the possibilities to assess the state of ecosystems, communities and populations according to the peculiarities of variation in abundance. The most accessible integral characteristic of populations is abundance species , with which many other parameters are closely related. To test the model, a short-term prediction of the number of anthophilous insects in the dry meadows of West Podillia was used.For each combination of parameters, the SARIMAX function from the statsmodels module is used, which selects the seasonal ARIMA model and assesses its overall quality. In this paper, we tried to predict the population size under conditions of uneven distribution of species and resources, as well as conduct a numerical study of possible scenarios for the existence of a species in a given time interval (2000 - 2017 years). A description of the criteria used to create the optimal model and verify its correct operation. Considerable attention is paid to the implementation of the time series algorithm using the Python 3 programming language. Seasonal ARIMA (p, d, q) (P, D, Q) s is used to take into account seasonality. Here (p, d, q) are the non-seasonal parameters described above, and (P, D, Q) are similar parameters applied to the seasonal component of the time series. The parameter s determines the frequency of the time series. The main thing in the selection of time series data in the seasonal ARIMA model is to find the ARIMA value (p, d, q) (P, D, Q) s, which select the best parameter. The ARIMA model based on archival data can be an important tool for monitoring and predicting biodiversity of both insects and plants in areas with similar abiotic and biotic factors.
Keywords: anthophilous insects, entomophilous plants, population, time series, ARIMA model, prediction, SARIMAX function