Research On The Method Of Summarizing Massive Meteorological And Oceanographic Non-hierarchical Spatial Data

With the continuous improvement of the degree of refinement of numerical forecasts in the field of meteorology and ocean,the amount of meteorology and ocean data is increasing,which brings huge challenges to the storage and query of data.The data summarization method is a process of lossy compression of data.However,most of the existing data summarization methods are aimed at hierarchical data,and are not suitable for a large number of non-hierarchical data in the meteorological and marine fields.This paper conducts a data summary study on non-hierarchical spatial data of meteorology and ocean to improve the efficiency of data storage and query.The main research work and innovations of this paper are as follows.(1)This paper proposes and formalizes the problem of Non-Hierarchical Data Summaries in a two-dimensional data space for the first time.Furthermore,this paper proves that the problem is NP hard by mathematical methods,so it cannot be solved in polynomial time.Aiming at the approximate solution of this problem,this paper designs the MS(Mark and Select)algorithm.The algorithm completes the approximate solution of the NHDS problem through two key steps: identifying eligible rectangles and selecting from the rectangle set to cover the data set.In theory,this paper proves the upper boundary of the approximate solution of the MS algorithm relative to the optimal solution.In practice,this article sets up related experiments to compare the performance of the MS algorithm and the current advanced data summary algorithm CA algorithm on the three data sets of global temperature data,humidity data,and synthetic data.The experimental results show that whether it is in running time,or in the worst error,the MS algorithm is better than the CA algorithm.(2)Aiming at the above-mentioned MS algorithm running efficiency is not high enough,the calculation object is limited to the two-dimensional data space,this paper has carried out the MS algorithm optimization and algorithm expansion work.In terms of algorithm optimization,this paper uses the heap algorithm to optimize,which can significantly reduce the amount of calculation in the score calculation process.Related experimental results show that the execution efficiency of the MS algorithm is improved by an order of magnitude after using the heap algorithm.In terms of algorithm expansion,this article extends the MS algorithm to three-dimensional space,and examines the execution efficiency and compression effect of the MS algorithm in the three-dimensional space.Experiments show that the scale of the MS algorithm abstract in the three-dimensional space is proportional to the amount of data,and the execution time increases super-linearly with the amount of data.