ABSTRACT
This chapter presents the estimation of the electrotonic parameters of the shunt cable model by three different methods. Linear Associative Memory (LAM) parameter estimation consists of two phases: a training phase and a recall phase. The iterative algorithm for the implementation of the weighted linear associative memory (WLAM) and procedures for avoiding local minima have been implemented and tested for a highly nonlinear model as well as the somatic shunt cable model. Error minimization techniques are currently the method of choice in most parameter estimation problems. However, it is also clear from these studies that LAM is as accurate as the error minimization method when small amount of noise are present in the data. Both aspects in WLAM contribute to a more efficient search and produce more accurate results than those obtained with LAM in two different estimation problems, the Van der Pol equation and the somatic shunt cable model.
