CMStatistics 2021: Start Registration
View Submission - CMStatistics
B0228
Title: Predicting inner temperature, humidity, and weight of beehives by using VAR models and sensorization Authors:  Carlos Javier Perez Sanchez - University of Extremadura (Spain)
M Isabel Parra Arevalo - Universidad de Extremadura (Spain)
Maria del Carmen Robustillo Carmona - Universidad de Extremadura (Spain) [presenting]
Abstract: Bees play an important role in both agriculture and the environment. Precision beekeeping is a useful tool for predicting the state of a hive, which can anticipate different events such as hive collapse, swarming or the disease presence. The objective is to predict the hive's weight, temperature, and humidity using sensor data and meteorological information. For this purpose, data obtained by the we4bee project in the hive of Grund-und Mitteschule Vohburg have been used. The studied dataset collects information about internal temperature and humidity, weight, external weather conditions and some information provided by the beekeeper. VAR models were considered to predict the internal state of the beehive (temperature, humidity, and weight) employing historical data series that include exogenous meteorological data. Predictions were made for one, three and seven days ahead. To validate this model, a 100-fold cross-validation was performed, obtaining a mean absolute error (mean +- standard deviation) of 0.156 +- 0.137 kg in weight predictions, 0.987 +- 0.663 C in temperature and 2.925 +- 2.437 \% in humidity. Given the promising results, we establish a starting point for predicting the state of the hives, which has not been addressed enough up to now.