Research on accurate recognition of local outlier data in cloud computing database

In view of the low accuracy and low efficiency of traditional methods, a new method for outlier recognition is proposed based on the combination of grid unit and density. The local outlier recognition criteria are introduced. According to the pruning rule, the set of points in the cloud database that cannot be local outliers is deleted in advance, and the remaining candidate local outliers are analyzed, and the data points satisfying the conditions are selected as local outliers. Meshing is performed to divide the data in the cloud computing database into a number of disjoint small rectangular cell areas. Rough screening and fine screening are used to determine whether the data is a local outlier based on the density of these cell areas, the specified thresholds, and the data of neighboring cells. When the cloud computing database changes, the corresponding processing rules are given. The processing code is given when the cloud computing database distribution includes or intersects. The proposed method is parallelized by MapReduce to ensure the efficient operation of the method. The simulation data experiment and network intrusion data experiment verify the accuracy and efficiency of the proposed method. © 2020 Editura Politechnica. All rights reserved.