在微细胞检测过程中，白细胞检出是白细胞自动识别系统中最基本也是最为关键的一个环节，其准确性和稳定性直接影响到整个系统的运行速度和识别准确率。由于细胞图象的复杂性，细胞的检出仍然是活跃的研究领域。文章重点研究了白细胞检测方法，在现行方法的基础上，提出了一种基于模糊形态学方法的新算法用于白细胞检测，实验结果表明这种方法比其它方法有较好的优势。

癌基因表达数据集具有小样本、高维数之特点，一般的机器学习机难以对其有效分类。因此，通常需要采用某些特征提取度量标准来进行降维处理。可是常用的一些特征提取度量标准亦会导致分类效果欠佳之问题。依据微分容量控制学习机DEEM，提出了一个新的特征提取度量标准NFEC，然后依据NFFC和DCCM，提出了适于癌基因表达数据集的特征提取算法DECFE。实验表明，新的度量NFEC和新的特征提取算法DCCFE较之现有方法对癌基因表达数据集分类时更为有效。本文的工作意义在于：（1）提出了一个新的更有意义的特征提取度量标准；（2）DCCM可以采用比核函数更为一般的一阶可微函数。因而提出的新的特征提取算法更具普遍应用意义。

In this paper, a novel robust maximum entropy clustering algorithm RMEC, as the improved version of the maximum entropy algorithm MEC [2-4], is presented to overcome MEC's drawbacks: very sensitive to outliers and uneasy to label them. Algorithm RMEC incorporates Vapnik's ε-insensitive loss function and the new concept of weight factors into its objective function and consequently, its new update rules are derived according to the Lagrangian optimization theory. Compared with algorithm MEC, the main contributions of algorithmRMECexit in its much better robustness to outliers and the fact that it can effectively label outliers in the dataset using the obtained weight factors. Our experimental results demonstrate its superior performance in enhancing the robustness and labeling outliers in the dataset.

针对模糊聚类神经网络FCNN（fuzzy clustering neural network）对例外点（outliers）敏感的缺陷，通过引入Vapnik's ε-不敏感损失函数，重新构造网络的目标函数，并根据拉格朗日优化理论推导出新的鲁棒模糊聚类神经网络及其算法（robust fuzzy clustering neural networks,简称RFCNN）。RFCNN有效地克服了FCNN对例外点敏感之缺点并且能得到合理的聚类中心。仿真实验结果表明，RFCNN较之于FCNN有更好的鲁棒性。

In this paper, a new selective feedback fuzzy neural network (SFNN) based on interval type-2 fuzzy logic systems is introduced by partitioning input and output spaces and based upon which a new FLS filter is further studied. The experimental results demonstrate that this new FLS filter outperforms other filters (e.g. the mean filter and the Wiener filter) in suppressing Gaussian noise and maintaining the original structure of an image.

Due to their universal approximation, fuzzy system with B-spline membership functions and CMAC neural network with B-spline basis functions have been extensively used in control. In many practical applications, they are desired to approximate not only the assigned smooth function as well as its derivatives. In this paper, by designing a fuzzy system and CMAC neural network with B-spline basis functions, we prove that such a fuzzy system and CMAC can universally approximate a smooth function and its derivatives, i.e, for a given accuracy, we can approximate an arbitrary smooth function by such a fuzzy system and CMAC that not only the function is approximate within this accuracy, but its derivatives are approximated as well. The conclusions here provide solid theoretical foundation for their extensive applications.

With the evidence framework, the regularized linear regression model can be explained as the corresponding MAP problem in this paper, and the general dependency relationships that the optimal parameters in this model with noisy input should follow is then derived. The support vector regression machines Huber-SVR and Norm-r r-SVR are two typical examples of this model and their optimal parameter choices are paid particular attention. It turns out that with the existence of the typical Gaussian noisy input, the parameter μ in Huber-SVR has the linear dependency with the input noise, and the parameter r in the r-SVR has the inversely proportional to the input noise. The theoretical results here will be helpful for us to apply kernel-based regression techniques effectively in practical applications.

提出了广义模糊CMAC（cerebellar model articulation controller）神经网络，并导出了其学习的充分条件。最后，证明了广义模糊CMAC在平方误差意义下的学习收敛性。研究结果为广义模糊CMAC的广泛应用提供了基础。