الفهرس | Only 14 pages are availabe for public view |
Abstract Our proposed erythemato-squamous Rough-Neuro model combines and optimizes first a rough set JohnsonReducer algorithm, second the neural network Multi-Layer Perceptron Levenberg-Marquardt classifier. JohnsonReducer algorithm reduces our 33 attributes to only 8 (76% reduction in complexity). This reduces the number of Levenberg-Marquardt classifier input acheving a classification error is at least comparable, if not smaller, when the set of inputs reduced Simulation results show that using the proposed model as an erythemato-squamous diseases system has a great impact on improving storage and time compared to other models. |