Data preparation for prediction and anomaly detection in precision beekeeping
Résumé
Faced with a major decline in the population of a variety of pollinators essential to our food security, it is essential to study their behavior and their evolution. The proposed approach is illustrated as an example of application in the beekeeping sector. Indeed, this sector is affected by a crisis due to the disappearance of bees and the different actors need support to make their decisions. We hope to offer new tools to help beekeepers predict the evolution of their hives, as well as tools to detect any anomalies. To meet these challenges, we have been looking for usable hive data. We have found data on line over a long period, but it needs to be completed and resampled. The dataset covers weight, internal humidity and internal temperature. In this article, we describe how these data have been prepared for prediction and anomaly detection by adding the external data like external temperature. Time-series resampling and interpolation techniques were used to help in this process. At the end of this process, the datasets are structured, cleaned and completed: they are ready to use in intelligent models.
Mots clés
Domaines
Informatique [cs]Origine | Publication financée par une institution |
---|