For the functional series model with a binary value (±1) as the input, the kernel, or the model parameters, can be rapidly estimated by using the fast Hadamard transform and the M sequence generated by a shift register as the input. Now the method is widely used to identify and analyze biological systems. As the first step in this paper, the case is considered in which the effect of the nonlinear relationship among the inputs on the system output must be examined, and it is pointed out that the physical interpretation of the kernel is easier when using the functional series model, whose input is {0, 1}, than when using the functional series model with {±1} assigned to the input. Then, the effective procedure is to identify the system using the functional series model, with {±1} as the input by a technique using the M sequence for rapidly estimating the kernel, after which the result is converted to a functional series model, whose input is {0, 1}, for interpreting the kernel. However, as yet there is no practical evaluative method for kernel estimating accuracy in the functional series model, whose input is {0, 1}. This paper notes that the kernel estimating accuracy in the functional series model whose input is {0, 1} is the same as that of the case where the kernel of the functional series model, whose input is {± 1}, is estimated and the result is converted to the functional series model whose input is {0, 1}. A simple evaluative method for kernel estimating accuracy is proposed based on this idea for the functional series model whose input is {0, 1}.