• Laser & Optoelectronics Progress
  • Vol. 57, Issue 21, 212001 (2020)
Xu Jian, Chen Qianqian*, and Liu Xiuping
Author Affiliations
  • 西安工程大学电子信息学院, 陕西 西安 710048
  • show less
    DOI: 10.3788/LOP57.212001 Cite this Article Set citation alerts
    Xu Jian, Chen Qianqian, Liu Xiuping. Classification of Electroencephalography Based on BP Neural Network Optimized By Crossover Operation of Artificial Bee Colonies[J]. Laser & Optoelectronics Progress, 2020, 57(21): 212001 Copy Citation Text show less
    Different kinds of individuals. (a) P1; (b) P2
    Fig. 1. Different kinds of individuals. (a) P1; (b) P2
    Schematic of crossover operation
    Fig. 2. Schematic of crossover operation
    Topological structure of three-layer BP neural network
    Fig. 3. Topological structure of three-layer BP neural network
    Improved global ABC to optimize process of BP neural network
    Fig. 4. Improved global ABC to optimize process of BP neural network
    Average fitness curves of three kinds of neural networks
    Fig. 5. Average fitness curves of three kinds of neural networks
    Prediction error curves of two networks
    Fig. 6. Prediction error curves of two networks
    Training errors of three kinds of neural networks
    Fig. 7. Training errors of three kinds of neural networks
    Comparison of accuracy between CGABC-BP and other classification methods
    Fig. 8. Comparison of accuracy between CGABC-BP and other classification methods
    TrialfunctionFunctionnameFunction expressionSearch scopeOptimalvalue
    Griewankf5(r)=14000i=1Qri2-i=1Qcosri i +1[-10,10]f5(0)=0
    Rastriginf2(r)=i=1Q[ri2-10cos(2πri)+10][-5.12,5.12]f2(0)=0
    Spheref1(r)=i=1Qri2[-5.12,5.12]f1(0)=0
    Rosenbrockf6(r)=i=1Q-1[100(ri+1-ri2)2+(ri-1)2][-2.048,2.048]f6(0)=0
    Table 1. Standard test functions
    Trial functionMethodOptimal valueConvergence rate /sMeanStandard value
    SABC3.83×10-5900.93.88×10-31.04×10-16
    GABC0230.08.30×10-131.41×10-17
    CGABC0191.45.47×10-160
    SABC0122.102.23×10-14
    GABC023.200
    CGABC015.200
    SABC1.05×10-1528.75.20×10-159.38×10-17
    GABC3.10×10-1817.28.03×10-135.08×10-17
    CGABC9.43×10-1814.34.31×10-182.66×10-17
    SABC3.10×10-15752.34.20×10-150.98
    GABC9.10×10-1622.35.45×10-150.61
    CGABC1.03×10-1819.18.76×10-170.11
    Table 2. Performance comparison of three algorithms
    Neural networkMax iterationsMin interationsAverage interationsTarget error
    BP134657732964110-4
    GABC-BP32116421510-4
    CGABC-BP2146117510-4
    Table 3. Training results of different BP neural networks
    Neural networkP1P2P3P4P5P6P7Average accuracy
    BP81.284.583.785.782.483.282.983.4
    GABC-BP86.385.488.187.386.887.484.786.6
    CGABC-BP90.789.593.992.091.190.692.591.5
    Table 4. Accuracy results of feature classification of different BP neural networks unit:%
    Xu Jian, Chen Qianqian, Liu Xiuping. Classification of Electroencephalography Based on BP Neural Network Optimized By Crossover Operation of Artificial Bee Colonies[J]. Laser & Optoelectronics Progress, 2020, 57(21): 212001
    Download Citation