Research On Parallel Convolutional Neural Network Algorithm Based On Big Data

With the rapid development of Internet technology and the advent of the era of big data,makes big data compared with traditional data,with the characteristics of 4 v-mass,fast changing,modal,the total value is high,4 v characteristics in traditional classification algorithm and it is difficult to deal with large data processing platform,in recent years,parallel technology and the development of feature wide choices classification algorithm for large data processing provides a new perspective.DCNN is an important class of classification algorithms with strong feature selection,generalization and function approximation capabilities,and is widely used in image analysis,speech recognition,target detection,semantic segmentation,face recognition,automatic driving and other fields.Therefore,the research of DCNN based on big data has become the research hotspot of classification algorithm.Although deep learning technology represented by DCNN has made many important breakthroughs in the field of big data classification in recent years,there are still the following problems :(1)How to reduce the redundant parameters of deep convolutional neural network while ensuring classification performance?(2)How to further deal with the problem of slow convergence rate of the optimization function of DCNN algorithm?(3)How to achieve fast and uniform grouping of data,so as to improve the parallelism efficiency of the cluster?To solve the above problems,based on the research and analysis of parallel DCNN algorithm and mining efficiency and other relevant knowledge,this paper proposes two parallel DCNN algorithms :(1)PDCNNO,the optimization algorithm of parallel deep convolutional neural network;(2)Deep convolutional neural network MR-FPDCNN based on feature graph under big data.The main research work of these two parallel DCNN algorithms is as follows:(1)Aiming at the problems of too many redundant parameters,slow convergence speed and low Parallel efficiency of Parallel DCNN algorithm in big data environment,this paper proposes a Parallel Deep convolutional neural network optimization algorithm named PDCNNO.Firstly,the algorithm designs PFM strategy,trains network,and obtains the compressed network,which effectively reduces redundant parameters and reduces the time and space complexity of DCNN training.Secondly,a CGMSE was designed to obtain local classification results,which realized rapid convergence of Conjugate gradient method and improved the convergence speed of the network.Finally,in the Reduce phase,a LBRLAwas proposed to obtain the global classification results,which realized the fast and uniform grouping of data and improved the acceleration ratio of the parallel system.Experiments show that the algorithm not only reduces the time and space complexity of DCNN training in the big data environment,but also improves the parallelization performance of the parallel system.(2)Aiming at problems such as excessive network redundant parameters,poor parameter optimization ability and low parallel efficiency exist in DCNN algorithm under big data environment.In this paper,deep convolutional neural network algorithm based on feature graph and parallel computational entropy is proposed.The algorithm is MR-FPDCNN.First,the algorithm designed the FMPTL and the pre-training network to obtain the compressed DCNN,which effectively reduced the redundant parameters and reduced the computational cost of DCNN training.Secondly,the paper proposes the IFAS based on ISS,initializes DCNN parameters according to The "IFAS" algorithm,realizes the parallelization training of DCNN,and improves the optimization ability of network.Finally,in the Reduce phase,a DLBPCE is proposed to obtain global training results,realizing fast uniform grouping of data and increasing the acceleration ratio of the parallel system.Experiments show that this algorithm not only reduces the computational cost of DCNN training in big data environment,but also improves the parallelization performance of parallel system.