Reduction Algorithm For Skyline Query Results Based On Dimension Preferences

In recent years, the methods of data collection are more and more various, mass storage is becoming more and more popular. Thus, Skyline query was introduced into the field of database. Compared with other algorithms, Skyline query can reflect the overall outline of the target data set, resulting in lot of applications in the fields of multi-objective decision making, etc.When the dimensionality increases, the Skyline query result is proportional to the number of growth, this phenomenon is called the curse of dimensionality. In the context of high-dimensional database, using the traditional Skyline query alone cannot provide effective decision support. So refining the selection of tuples, according to a certain standard to streamline Skyline query result is very necessary.First of all, this paper introduces the problem of the curse of dimensionality and existing Skyline query results concise algorithm, and a brief analysis. Then, introduced the concept of user preference of dimension, the Skyline query results reduction algorithm with successive relaxation of control relationship based on the dimension preferences is proposed and introduced in detail. In the algorithm, the user preferences of dimension are represented as user’s qualitative sort of dimension of data sets. Algorithm using the indifference threshold given by users starts to relax the control relationship from the least important dimension, after each round has a more important dimension to be relaxed, until the algorithm returns the result to meet users’needs or the all the dimensions are relaxed. It also proposed several methods to optimize the comparison process of the algorithm, reduce the running time. After that, the algorithm is analyzed theoretically to prove the optimization methods. Finally, there were several different targeted experiments to prove the results and performance of the algorithm. Experiment proves that the algorithm can achieve the purpose of streamlining the Skyline result set to elect Skyline results meet more to customer needs. Results in the selection of the quality and performance of the algorithm are superior to comparative algorithms.