2020/12/07 by Bashar Alhnaity, Alhnaity, Bashar, Stefanos Kollias +9 · 1 citation
Agricultural and Biological Sciences · Chemistry · #FOS: Computer and information sciences #Greenhouse Technology and Climate Control #Leaf Properties and Growth Measurement #Machine Learning (cs.LG) #Spectroscopy and Chemometric Analyses
paper · pdf · doi:10.48550/arxiv.2012.04041
openalex publication_date 2020/12/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Multi-step prediction is considered of major significance for time series\nanalysis in many real life problems. Existing methods mainly focus on\none-step-ahead forecasting, since multiple step forecasting generally fails due\nto accumulation of prediction errors. This paper presents a novel approach for\npredicting plant growth in agriculture, focusing on prediction of plant Stem\nDiameter Variations (SDV). The proposed approach consists of three main steps.\nAt first, wavelet decomposition is applied to the original data, as to\nfacilitate model fitting and reduce noise in them. Then an encoder-decoder\nframework is developed using Long Short Term Memory (LSTM) and used for\nappropriate feature extraction from the data. Finally, a recurrent neural\nnetwork including LSTM and an attention mechanism is proposed for modelling\nlong-term dependencies in the time series data. Experimental results are\npresented which illustrate the good performance of the proposed approach and\nthat it significantly outperforms the existing models, in terms of error\ncriteria such as RMSE, MAE and MAPE.\n