An index-based classification scheme using neural networks for multiclass problems

This paper proposes a novel classification scheme based on a semi-supervised back-propagation (SSBP) learning algorithm for multiclass problems. The proposed approach can derive a fuzzy index as a classification quantifier for each specific class by means of a specially-defined cost function. Misclassifications can be removed through introducing an extra indeterminate class for some complicated non-probabilistic classification problems. The reliability of the classification results is improved basically as a result of creating the indeterminate class. Applications to a 3-pattern classification problem demonstrate the effectiveness of the proposed scheme.