Based on the recursive neural network in the FPGA implementation of LSSVM, selection of the initial-CodePudding

Mobile phone is not very good figure, edit later

CodePudding user response:

this is to look at the many recursive neural network is used to implement the LS - SVM training papers appear in the formula, is solving the Lagrange multiplier and b,

because of the need to use the FPGA implementation, the discretization equation, to get the two formulas above, because it is a recursive neural networks, we need to select an alpha generation into the calculation of initial values, then the how to determine the initial value, hope to get your bosses solutions,

below is in "I" image processing method based on FPGA and the SVM research found in validation of alpha and others b convergence figure, but not sure how much is the initial value,

CodePudding user response:

reference 1st floor Whistle0602 response:

this is to look at the many recursive neural network is used to implement the LS - SVM training papers appear in the formula, is solving the Lagrange multiplier and b,

because of the need to use the FPGA implementation, the discretization equation, to get the two formulas above, because it is a recursive neural networks, we need to select an alpha generation into the calculation of initial values, then the how to determine the initial value, hope to get your bosses solutions,

below is in "I" image processing method based on FPGA and the SVM research found in validation of alpha and others b convergence figure, but not sure how much is the initial value,

I also in the study of this problem, a blogger, 526250429 @qq.com.