论文部分内容阅读
为了解决未标识样本的分类问题,提出一种基于多维度收缩的、新的排序-模糊神经网络分类器模型OFMM.该模型首先利用多维度收缩法对输入的所有样本进行排序,然后获得样本间的相似性测度值.并利用该相似性测度值指导随后的分类器超盒扩张与压缩过程,从而使得该模型不仅提高对未标识样本进行有效分类的性能,而且无论是在网络结构方面,还是在训练时间方面都有所改进.有关标准数据集的实验结果表明,该模型明显优于传统的通用模糊神经网络,是一种较实用且有效的分类器.
In order to solve the classification problem of unidentified samples, a new multi-dimensional contraction-based OFMM, a sort-fuzzy neural network classifier model, is proposed. Firstly, the multi-dimensional shrinkage method is used to sort all input samples and then the sample space The similarity measure value is used to guide the subsequent classifier hypershelf expansion and compression process so that the model not only improves the performance of classifying the unidentified samples effectively but also improves the performance of the network in terms of network structure, The training time has been improved.Experimental results on the standard dataset show that the model is better than the traditional generalized fuzzy neural network is a more practical and effective classifier.