A framework for predicting the location of Vespa mandarinia nests via deep learning approach

Vespa mandarinia are capable of attacking other bees and humans and causing death. Thus, it is crucial to accurately identify their nest locations. In this paper, an autonomous update model based on deep learning is developed. The specific work includes three parts: (1) Establishing an ecological seed bank based on the biological characteristics of Asian hornets for the generation of scores of popular text information in the process of text sentiment analysis; (2) Determining the clustering categories by combining the recognition scores and Euclidean distances of image videos; and (3) Employing an artificial neural network model optimized by a greedy algorithm for risk identification. The empirical results show that the integrated prediction error probability of the method in this paper is 0.0439. Our research provides theoretical ideas and practical basis for effective early warning prevention and control deployment for biosecurity problems represented by animal invasion, and can reduce monitoring costs.