Research On Classification Based On Improved And Parallel Artificial Bee Colony Algorithm

In recent years,more and more experts and scholars have devoted themselves to the research of data classification.With the increasing number of data dimensions,the computational complexity and computation time of the classification algorithm will increase greatly.Traditional experience selection or large-scale grid search for model parameters can no longer meet the requirements.Therefore,this paper proposes a support vector machine parameters optimization algorithm based on artificial bee colony algorithm(ABC-SVM).The purpose is to solve the problem that the traditional methods to select parameters of the support vector machine is low efficiency.Firstly,using ABC to optimize the parameters of SVM,then the model is applied on UCI database.The results demonstrate the effectiveness and feasibility of ABC-SVM.The basic ABC uses one dimensional search strategy for food sources,which can lead to food source with better solutions in one dimension being discarded due to reaching the threshold of search,reducing the efficiency of the algorithm.Greedy search on all dimensions of food sources will greatly increase the computational complexity of the algorithm.For this reason,this paper proposes an improved multi-dimensional search strategy.First,find out the dimensions with update value,and then in-depth excavation of these dimensions.Using the standard test function to verify the performance of the algorithm,results show that IABC not only has good performance in improving convergence speed and calculation accuracy,but also reduces the time cost of the calculation.IABC is also used to optimize the parameters of the support vector machine.Using the same data set to test,it is found that IABC has a higher classification accuracy.Compared with MABC,IABC costs less time.Finally,in view of natural parallelism of ABC,combine parallel with ABC and use the test functions to verify the new algorithm.The result shows that ABC with parallel strategy is effective.Then the new algorithm is used for parameter optimization of the support vector machine and classify the human activity data.The experimental results show that ABC with parallel strategy can reduce the time cost.And MABC can improve the convergence speed and classification accuracy.After combining parallel with MABC,the improved algorithm has both advantages and broad application prospects.