Prediction of Plasma Enhanced Deposition Process Using GA-Optimized GRNN
A genetic algorithm (GA)-based optimization of generalized regression neural network (GRNN) was presented and evaluated with statistically characterized plasma deposition data. The film characteristics to model were deposition rate and positive charge density. Model performance was evaluated as a function of two training factors, the spread range and a factor employed for balancing training and prediction errors. For comparison, GRNN models were constructed as well as four types of statistical regression models. Compared to conventional GRNN models, GA-GRNN models improved the prediction accuracy considerably by about 50% for either film characteristic. The improvements over statistical regression models were more pronounced and they were more than 60%. There results clearly reveal that the presented technique can significantly improve conventional GRNN predictions.
KeywordsDeposition Rate Plasma Enhance Chemical Vapor Deposition Generalize Regression Neural Network Pattern Layer Film Characteristic
Unable to display preview. Download preview PDF.