beepnl / bee-health-calculation-model-bgood-wp5

Bee health data calculation model B-GOOD Work package 5

bee-health-calculation-model-bgood-wp5

Bee health data calculation model B-GOOD Work package 5. This code was developed by SESS in Aarhus University with the support of the B-GOOD project: https://b-good-project.eu/.

Description

Calculates a colony survival prediction value for a single hive based on an array of hourly weight data. The input weight data is assumed to be the 'cleaned weight' (beekeeper actions removed from the weight measurements), evenly spaced with 1-hour intervals. Inference is ran on the input data using a pre-trained PyTorch LSTM model (hsi.pt).

Running the model

1. Install dependencies

• Python3: https://www.python.org/downloads/
• PyTorch: https://download.pytorch.org/whl/torch_stable.html
• Numpy: https://numpy.org/

2. Add your own data

Replace test_data with array of cleaned weight data

3. Run the model

python3 b_good_prediction.py

Author

Xiaodong Duan, [email protected]

Input data

Measurement

• net_weight_kg (weight excluding beekeeper actions)

Data interval

• 1 hour average

Data period

• relative interval of 30 * 24 hours, looking back from now

Data items

• data per hive

Output data

• float between 0 and 1