La inteligencia artificial en la predicción de la temperatura ambiental y del suelo en Ecuador

Introduction: The main objective of the study was to analyze the probability and prediction for environmental and soil temperature in the coastal area of Manabí in Ecuador. Methodology: The methodology makes use of Box Jenkins ARIMA time series and comparison of means. The data was measured at 07:00 am, 12:00 pm and 18:00 pm, starting in January 2015 until December 2020. The data was analyzed and processed with the help of artificial intelligence incorporated into the RStudio software. Results: The results show that soil temperature is correlated with environmental temperature. Discussions: Goodness-of-fit tests for the coefficients and assumptions validated the observed and expected ARIMA model. Furthermore, the AIC and BIC criteria were used to choose the best predictive model. Conclusions: In conclusion, artificial intelligence identified that the prediction of ambient and soil temperatures are adequately simulated through an ARIMA(0,1,1)(0,1,1)[12] model, with trend and seasonality components, By affirming a non-stationary time series model, it is determined that temperature has a small variability for each period of time, but increasing, and in the future this climatic factor will probably become a determinant of global warming.