Intelligent Spatial Interpolation-based Frost Prediction Methodology using Artificial Neural Networks with Limited Local Data
share
The weather phenomenon of frost poses great threats to agriculture. Since it damages the crops and plants from upstream of the supply chain, the potential impact of frosts is significant for agriculture-related industries. As recent frost prediction methods are based on on-site historical data and sensors, extra development and deployment time are required for data collection in any new site. The aim of this article is to eliminate the dependency on on-site historical data and sensors for frost prediction methods. In this article, a frost prediction method based on spatial interpolation is proposed. The models use climate data from existing weather stations, digital elevation models surveys, and normalized difference vegetation index data to estimate a target site's next hour minimum temperature. The proposed method utilizes ensemble learning to increase the model accuracy. Ensemble methods include averaging and weighted averaging. Climate datasets are obtained from 75 weather stations across New South Wales and Australian Capital Territory areas of Australia. The models are constructed with five-fold validation, splitting the weather stations into five testing dataset folds. For each fold, the other stations act as training datasets. After the models are constructed, three experiments are conducted. The first experiment compares the results generated by models between different folds. Then, the second experiment compares the accuracy of different methods. The final experiment reveals the effect of available stations on the proposed models. The results show that the proposed method reached a detection rate up to 92.55 alternative solution when on-site historical datasets are scarce.
READ FULL TEXT