A new training algorithm for the general regression neural network

Masters, Timothy; Land, Walker H.

doi:10.1109/icsmc.1997.635142

Cited by 59 publications

(35 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The DE convergence plots often exhibit a characteristic staircase behavior, with intervals where the cost does not decrease from generation to generation. To mitigate this drawback somewhat, we have implemented an ad hoc hybrid method, which combines DE with an occasional deterministic descent along the negative gradient of the cost function, with the gradients computed by the finite-difference method [34,35]. The descent is triggered with a probability of 0.5 each time the best cost does not decrease between consecutive generations.…”

Section: Numerical Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Electromagnetic Imaging of Dielectric Cylinders by Differential Evolution and Single Integral Equation

Michalski¹

2013

PIER B

View full text Add to dashboard Cite

Abstract-A technique is described for the electromagnetic reconstruction of the location, shape, dielectric constant, and conductivity of buried homogeneous cylinders of elliptic cross-section. The inversion procedure is based on the Differential Evolution algorithm and the forward problem is solved using the single boundary integral method. Simulation results are presented which demonstrate that this hybrid approach can offer a conceptually simple yet efficient and reasonably robust method for the imaging of buried objects and voids.

show abstract

Section: Numerical Resultsmentioning

confidence: 99%

“…It is perhaps worthwhile to point out some differences between the DE and traditional, real-coded genetic algorithms (GAs) [33,34,39]. First, DE does not involve selection of parents based on fitness.…”

Section: Appendix B Differential Evolutionmentioning

confidence: 99%

Electromagnetic Imaging of Dielectric Cylinders by Differential Evolution and Single Integral Equation

Michalski¹

2013

PIER B

View full text Add to dashboard Cite

show abstract

“…When WN is applied in the real word, for example, the morphology variations of ECG waveforms are different for different patients, and even for the same patient or for the same type (Osowski & Linh, 2001), traditional networks can become a bottleneck requiring retraining with new features added into the current database. PNN (Specht et al, 1988) and general regression neural networks (GRNN) (Masters & Land, 1997;Seng et al, 2002) have been presented, and are recognized as having expandable or reducible network structure, fast learning speed, and promising results. In these adaptation methods, the choice of smoothing parameter has significant effects on the network outcome, and the choice of parameter is usually based on the overall statistical calculation from pre-collected training data.…”

Section: Introductionmentioning

confidence: 99%

Adaptive wavelet network for multiple cardiac arrhythmias recognition

Lin

Chen

2008

Expert Systems with Applications

View full text Add to dashboard Cite

“…Advantages of these approaches are shown in (Storn & Price, 1997). Using DE for training neural networks was first introduced in (Masters & Land, 1997). It was reported that the DE algorithm is particularly suitable for training general regression neural networks (GRNN), and it outperforms other training methods such as gradient and Hessian on applications which have the presence of multiple local minima in the error space.…”

Section: Introductionmentioning

confidence: 99%