The Labeled Classification and its Application

摘要: This paper presents and evaluates a new classification method that aims to improve classifiers performances speed up their training process. The proposed approach, called labeled classification, seeks convergence of the BP (Back propagation) algorithm through addition an extra feature (labels) all examples. To classify every example, tests will be carried out each label. simplicity implementation is main advantage this approach because no modifications are required in algorithms. Therefore, it can used with others techniques acceleration stabilization. In work, two models proposed: LMLP (Labeled Multi Layered Perceptron) LNFC Neuro Fuzzy Classifier). These tested using Iris, wine, texture human thigh databases evaluate performances.