Data Mining to Characterize Seasonal Patterns of Apis mellifera Honey Bee Colonies

Abstract

Among the agricultural crops used for human consumption, 75% depends on pollination. As the principal pollinating agent, bees are essential for the food production for humans and the ecosystems sustainability. However, a combination of habitat destruction, climate change and exposure to pesticides and pathogens has led to a significant decrease in bee population. Here we propose a method to recognize status patterns of Apis mellifera colonies through the application of data mining techniques. Using a real dataset from the HiveTool.net containing Apis mellifera temperature, humidity and weight data, we identified 3 status patterns in the observed hive. Our results suggest that the recognized patterns are consistent with a honey bee colony life cycle. Based on the found patterns, we propose a high accuracy classification model capable of automatically identifying colony status for new samples.