# Honeybee Colony Growth Period Recognition Based on Multivariate Temperature Feature Extraction and Machine Learning

## TL;DR

This paper introduces a machine learning method to identify honeybee colony growth periods using temperature data, improving decision-making for beekeepers.

## Contribution

A novel approach combining multivariate temperature features and machine learning for intelligent recognition of honeybee colony growth periods.

## Key findings

- The proposed method effectively characterizes honeybee colony growth periods using temperature data.
- BP method achieved the best performance with a mean absolute error of 1.45%.
- The method proved practical across different regions, showing adaptability.

## Abstract

Identifying the growth period of bee colonies can guide beekeepers to make better decisions and promote the development of bee colonies. Unlike traditional manual experience-based recognition, this paper proposes a new approach, which combines multivariate temperature feature extraction and machine learning to intelligently recognize the growth period of bee colonies. Firstly, the year-round temperature data from 38 hives in Tai’an and Guilin was collected. Then, the 17 time domain characteristic indices were extracted from this dataset. To acquire the most sensitive features, the impact of different time scales on temperature feature extraction was analyzed. Subsequently, principal component analysis (PCA) was employed to reduce the dimensionality of the original feature vectors, thereby decreasing computational load and enhancing feature sensitivity. Finally, six machine learning algorithms, including both supervised and unsupervised learning, were utilized to identify the growth period of bee colonies. The results demonstrate that the proposed features can effectively characterize the growth period of bee colonies, and the BP method performs best in predicting growth period categories, with an MAE of only 1.45%. Moreover, the identification results of different regions also prove the practicability of the proposed method.