Temporal and Spatial Evolution Pattern of Terrestrial Ecosystem based on Machine Learning Algorithm

With the continuous development of global terrestrial ecosystem(TE) research, a large number of relevant observation data have been accumulated. In the era of big data, with the continuous innovation of machine learning theory and technology, these new methods provide unprecedented opportunities for the analysis of temporal and spatial evolution patterns of TEs. This paper analyzes the temporal and spatial evolution pattern of TE based on machine learning algorithm(MLA), discusses the machine learning method and its application in the era of big data, studies and analyzes the temporal and spatial evolution of ecosystem service value and the accounting of temporal and spatial evolution pattern of TE; MLA is proposed, and six different machine learning methods are tested through experiments to evaluate the daily GPP flux in the model prediction stage. The experimental results verify the effectiveness and feasibility of the spatial-temporal evolution pattern analysis of TE based on MLA proposed in this paper.