Radial Basis Function Neural Network Based on Genetic Algorithms for Evaluating Cigarettes Aroma

The aroma quality of cigarettes is often evaluated by sense of experts in quality of cigarettes. The accuracy of evaluation results is hard to be guaranteed, so it is necessary to research accurate and objective evaluation method now. In this paper, the evaluation method on aroma of cigarettes is deeply researched using Radial Basis Function (RBF) neural network based on genetic algorithms and smell sensor array. Experimental results show clearly that: this method is feasible, the learning algorithm given by this paper for RBF neural network is effective. Some foundations for going further into the objective evaluation method of cigarettes aroma are laid.