Research On Skyline Query Based On Spatial Data With Numeric And Nominal Attributes

In this paper, we aim to study the materialization algorithm for dynamic skyline query on spatial data with numeric and nominal attributes. Based on the in-depth analysis on the research status at home and abroad, we focus on the property of nominal attribute on which different users have different preferences on the attribute values and make a study on the method for skyline query on spatial data with numeric and nominal attributes.Firstly, the existing skyline dominant analysis methods can only be used in a single space of spatial data with numeric and nominal attributes. However, there are the real-time demands for responding dynamic preference skyline query on any subspace of spatial data with numeric and nominal attributes by users. In order to satisfy the needs, a sharing strategy for skyline dominant analysis conditions between different subspaces is proposed. On the basis of this sharing strategy, a corresponding algorithm is also presented to materialize all subpaces skyline dominant analysis conditions in a space.Secondly, it is analysed that the existing methods can merely materialize the dynamic preference skyline query on a single space of spatial data with numeric and nominal attributes. For the existing materialization method can only compute the skycube of spatial data with numeric attributes, a materialization method, which is called semi-materialization skycube, is employed to perform all the dynamic preference skyline queries on all subspaces of spatial data with numeric and nominal attributes. Moreover, combining the proposed algorithm which is based on the sharing strategy for skyline dominant analysis conditions between different subspaces, we present a materialization skycube algorithm to compute the semi-materialization skycube by improving the sharing strategy of different subpaces skyline of spatial data with numeric and nominal attributes.Thirdly, considering the characteristics of dynamic preference skyline query on semi-materialization skycube and the shortcoming of linear table that stores the semi-materialization skycube, we design an index structure, which improve the enumeration tree, to store semi-materialization skycube. In addition, according to the feature of this index structure, a skyline query algorithm is given to handle dynamic preference skyline query on any subspace by shortest path stored in the index structure.Finally, by conducting contrast experiments respectively on two different datasets which are called anti-correlated database and independent database, we compare our proposed algorithms with the na?ve algorithms on the aspect of time efficiency.