一种用于模式分类有监督的模糊ART神经网络

A Supervised Fuzzy ART Neural Network for Pattern Classification

探讨了一种将有监督学习机制融合到模糊ART网络构成一个有监督的模糊ART神经网络模型．这种网络能同时处理有监督和无监督学习问题，并具有积累和增加网络学习的能力．对该网络进行了滚动轴承检测数据模式分类实验，并与BP网络进行了比较性实验．结果表明：该网络具有良好模式分类能力和较好的可塑性．

A new neural network model that incorporates a supervised mechanism into a fuzzy ART is investigated. The model can cope with supervised learning and unsupervised learning simultaneously, and has the ability of incremental learning. A few experiments of bearing pattern classification prove pefformance of the model and by comparing pefformance of the model with BP model. The results of experiments indicate tha the model has the ability of pattem classification and flexibility.