Research on Method of Optimizing Mining Method of Abnormal Data in Parallel Database

Considering semantic feature clustering closely and separation of parallel abnormal data in the database optimization classification algorithm is presented in this paper. Constructing parallel database anomaly data semantic feature separability measure model is obained, design semantic mapping network structure to achieve noise of abnormal data preliminary separation, design semantic features clustering compactness and separation algorithm, to contain noise and outliers in the parallel database system for interference suppression, mining classification of abnormal database optimization. Simulation results show that the algorithm improves the parallel database abnormal data search the mining process of precision and classification mining compactness and high accuracy.