A New Shelf Life Prediction Method for Farm Products Based on an Agricultural IOT

Abstract

Shelf life prediction is an important problem in the field of food safety. This problem has been extensively studied in different research fields. In this paper, based on our agricultural IOT (Internet of Things) platform, we study this problem from the viewpoint of data mining. In our agricultural IOT platform, by setting various types of sensors, it is possible for us to collect information of a farm product during its whole life cycle such as planting, storage, processing, transportation and sale. Shelf life of a farm product is very difficult to determine since it will be affected by many factors during its life cycle, such as temperature, air/soil humidity etc. After integrate raw sensor data streams into batch id-based data streams, we adapt Back-Propagation method to the integrated sensor data streams to predict the shelf life of a farm product. Experiments are conducted on real data from our agricultural IOT platform and the experimental results demonstrate that the proposed method could provide a very good prediction for the shelf life of a farm product.