A Parallel Minimum Attribute Co-reduction Accelerator based on Quantum-inspired SFLA and MapReduce Framework

The fast increase and update of big data brings a new challenge to quickly acquire the useful information with classical attribute reduction methods. In this paper, a parallel minimum attribute co-reduction accelerator (QSMFAC) based on quantum-inspired SFLA and MapReduce framework is presented. First, a novel framework of N-populations distributed co-evolutionary cloud model is designed to divide the entire population into N subpopulations and share the rewards of different subpopulations’ solutions under MapReduce mechanism. Second, the divided attribute subsets in subpopulations are coevolved by quantum-inspired SFLA in which evolutionary frogs are represented by quantum chromosome gene state, and the crossover co-evolutionary strategy between neighborhood subpopulations can adapt the consecutive sharing of better performance. Third, the MapReduce based approximation parallelism mechanism is adopted to conduct rules reduction to speed up the computation of attribute equivalence classes, so that it will be extended to high performance in both quality of solution and competitive computation complexity. Experimental results indicate the proposed accelerator has better on efficiency and accuracy of minimum attribute reduction than some representative algorithms. Moreover it is applied into MRI segmentation with intensity inhomogeneity, and the effective and robust segmentation results further indicate it has stronger superior for complex big data application.