The Study Of Forest Ecological Station Data Clustering Based On Big Data

Huge amount of data have been accumulated in the past several decades in Long Standing Observation Station of Forest Ecosystem (or "Forest Ecosystem Station", "FEC" for short) of China. In recent years, researchers both domestically and abroad mine the FEC data mostly based on sampling and statistical analysis of small amount of temporal and spatial data. It is of great importance to clustering the whole FEC data, since it can help promote the development of digital forestry.Firstly, the paper introduces the research situation and significance of the FEC data clustering, and then it describes the significance, principles and common practices of current big data processing. Secondly, clustering system on FEC data is constructed based on big data processing. The knowledge and principles are described in detail about partition clustering, hierarchical clustering based on the covering relations between tree nodes, and Hadoop distributed systems. It is explained why the old hierarchical clustering algorithm cannot be adapted to distributed computing systems. Thirdly, based on the above analysis, a new hierarchical clustering algorithm, DHCSA, is proposed and is applied in data analysis. After that, the functions of the clustering system of FEC data are described in detail. The last part concludes the paper and prospected the future work.The innovation aspect of this paper is that a new hierarchical clustering algorithm, DHCSA, is proposed which is based on distributed computing environment. DHSCA is able to significantly reduce computing time and it can run not only on a single-node computing environment but on a distributed computing environment.