A fuzzy controlled neural network for screening new product ideas
Yi‐Chen Lin, Jong-Mau Yeh · Journal of Information and Optimization Sciences · 2001
Success in product development is a critical management issue for the modern firm, especially those in technology driven industries. Opportunity identification is the initial stage in the new product development process where ideas for new product are generated and screened. One interesting area for the use of neural networks is in event prediction. This research develops a back-propagation neural network model for prediction of screening new product ideas and tests it using data from experts. A comparison of the predictive abilities of both the neural network and the discriminate analysis method is presented. On major drawbacks of the back-propagation model is long training time. This research will use fuzzy control theory to overcome it. The fuzzy control rules can control learning rate by the relation of the error and the sum of weight changes in one learning process, so that neural network can learn without the stagnation. This study applied fuzzy control theory to neural network to avoid the stagnation, so that the neural network can learn the system quickly to overcome its major drawbacks. This fuzzy controlled neural network will be used in screening new product ideas to improve the quality of new product development and the possibility of identifying successful new products.