Identifying Bee Species by Means of the Foraging Pattern Using Machine Learning
Author(s)
Arruda, H
Imperatriz-Fonseca, V
De Souza, P
Pessin, G
Griffith University Author(s)
Year published
2018
Metadata
Show full item recordAbstract
Bees are agents of nature that help provide about one third of the food we eat through a process called pollination. The primary aim of this work is to classify bee species and this was achieved by employing different feature vectors and machine learning algorithms for gathering foraging pattern data, with radio frequency electronic tags glued to the bees' thoraxes. Each time a bee entered or left the hive, the timestamp was stored. The data were analyzed in a time series format, in which the bees' activities were grouped into different categories. The Random Forest algorithm achieved the best results with the area under a ...
View more >Bees are agents of nature that help provide about one third of the food we eat through a process called pollination. The primary aim of this work is to classify bee species and this was achieved by employing different feature vectors and machine learning algorithms for gathering foraging pattern data, with radio frequency electronic tags glued to the bees' thoraxes. Each time a bee entered or left the hive, the timestamp was stored. The data were analyzed in a time series format, in which the bees' activities were grouped into different categories. The Random Forest algorithm achieved the best results with the area under a ROC curve of 0.94 and 87.41% degree of accuracy, by grouping 12 bees and using 72 attributes.
View less >
View more >Bees are agents of nature that help provide about one third of the food we eat through a process called pollination. The primary aim of this work is to classify bee species and this was achieved by employing different feature vectors and machine learning algorithms for gathering foraging pattern data, with radio frequency electronic tags glued to the bees' thoraxes. Each time a bee entered or left the hive, the timestamp was stored. The data were analyzed in a time series format, in which the bees' activities were grouped into different categories. The Random Forest algorithm achieved the best results with the area under a ROC curve of 0.94 and 87.41% degree of accuracy, by grouping 12 bees and using 72 attributes.
View less >
Conference Title
Proceedings of the 2018 International Joint Conference on Neural Networks (IJCNN)
Subject
Artificial Intelligence and Image Processing