An algorithm to compute composition skyline based on standard deviation

The most interesting information can be got by composition skyline query. Although this query can return the result as a group, the size of result usually is too large, so the practicability of this kind of query needs to be further strengthened. In order to reduce the size of result set, we propose an efficient algorithm by using standard deviation, which can return a smaller size of query set. Firstly, we introduce skyline layer and directed dominant graph to compute all the dominant relationships of all the points, then we use standard deviation to compute the dispersion of each group, at last the standard deviation is used as an attribute of group to compute composition skyline. We simulate three kinds of datasets to test our algorithm, and analyse the experimental result to shows the efficient of algorithm.