Honey bee colony population daily loss rate forecasting and an early warning method using temporal convolutional networks

Abstract

The population loss rate of a honey bee colony is a critical index to verify its health condi-tion. Forecasting models for the population loss rate of a honey bee colony can be an essential tool in honey bee health management and pave away to early warning methods in the understanding of potential abnormalities affecting a honey bee colony. This work presents a forecasting and early warning algorithm for the population daily loss rate of honey bee colonies and determining warning levels based on the predictions. Honey bee colony population daily loss rate data were obtained through embedded image systems to automatically monitor in real-time the in-and-out activity of honey bees at hive entrances. A forecasting model was trained based on temporal convolutional neural networks (TCN) to predict the following day’s population loss rate. The forecasting model was optimized by conducting feature importance analysis, feature selection, and hyperparameter optimization. A warning level determination method using an isolation forest algorithm was applied to classify the population daily loss rate as normal or abnormal. The integrated algorithm was tested on two population loss rate datasets collected from multiple honey bee colonies in a honey bee farm. The test results show that the forecasting model can achieve a weighted mean average percentage error (WMAPE) of 17.1 ± 1.6%, while the warning level determination method reached 90.0 ± 8.5% accuracy. The forecasting model developed through this study can be used to facilitate efficient management of honey bee colonies and prevent colony collapse.