Fast Clustering Optimization Method of Large-Scale Online Data Flow Based on Evolution Incentive

In order to optimize large-scale online data stream clustering quality and algorithm performance, a fast clustering optimization method is proposed based on line data flow evolution incentive. Firstly, in the online stage, the high density partition algorithm is used to generate micro cluster set, and the set takes the snapshot window as the time scale. The micro cluster set is updated in real-time. In the offline stage, the micro cluster set is read, and the improved quantum genetic algorithm and the evolution incentive function are used for the iterative optimization in clustering center. Finally, the adaptive mutation operator is taken as the optimal variation operation for the populations. The global search ability of the algorithm is improved. Simulation results show that algorithm has better clustering accuracy, and the convergence speed is high with less memory cost. It has good application value in practice.