一种电能质量多扰动分类中特征组合优化方法

针对电能质量扰动分类中冗余特征量造成分类器训练困难、分类准确率下降的问题,提出一种基于改进遗传算法的特征组合优化方法。该方法对信号进行小波变换,提取各层的改进小波能量熵作为原始特征,并构造一种基于欧氏距离的适应度函数,采用改进的自适应遗传算法对原始特征进行筛选和优化组合,形成用于电能质量扰动分类的最优特征组合。分别采用二分类-支持向量机法(Binary-SVM)、多标签径向基神经网络(ML-RBF)和多标签K近邻法(ML-KNN)对不同噪声情况下的电能质量单一扰动和混合扰动进行分类,仿真结果验证了所提特征组合优化方法能有效提高分类器的训练速度和分类准确率。     

The use of seed texture features for discriminating different cultivars of stored apples

The aim of this study was to identify the textural features of apple seeds with the highest discriminatory power for distinguishing the seeds of different apple cultivars with the use of discriminative classifiers. The seeds of apple cvs. Gala, Jonagold and Idared were scanned with the use of a flatbed scanner, and the acquired images were processed to calculate textural features from color channels: L, a, b, R, G, B, Y, U, V, H, S, I, X, Y and Z. The selected textures were used to develop discriminative models and distinguish the seeds of the examined apple cultivars. The analyses were performed for color spaces and color channels. The seeds of apple cvs. Gala and Idared were discriminated with 100% accuracy in models based on the textures from Lab and YUV color spaces and color channel L for the Naive Bayes, Multilayer Perceptron and Multi Class classifiers. The discriminatory accuracies of the seeds of all analyzed apple cultivars (Gala, Idared and Jonagold) ranged from 72% to 85%. The discriminatory accuracy of the textures selected from Lab color space for the Naive Bayes classifier reached 85%. The seeds of apple cvs. Gala and Jonagold were discriminated with 78–90% accuracy, and the discriminatory accuracy of the textures from Lab color space and color channel b for the Naive Bayes classifier reached 90%. The seeds of apple cvs. Idared and Jonagold were distinguished with 80–94% accuracy. The models based on textures from Lab color space and color channel b for the Naive Bayes classifier were characterized by 94% discriminatory accuracy. The study demonstrated that textural features are useful for discriminating the seeds of different apple cultivars.     

Soft computing methods applied to combination of one-class classifiers

The paper shows the possibilities of generalizing the two-class classification into multi-class classification by means of a fuzzy inference system. Fuzzy combiner harnesses the support values from classifiers to provide final response having no other restrictions on their structure. We compare proposed combination methods with ECOC and two variations of decision templates, based on Euclidean and symmetric distance. The effectiveness of the proposed combination method based on the fuzzy logic theory is also evaluated via computer experiments carried out on benchmark datasets.     

Evaluation of incremental learning algorithms for HMM in the recognition of alphanumeric characters

We present an evaluation of incremental learning algorithms for the estimation of hidden Markov model (HMM) parameters. The main goal is to investigate incremental learning algorithms that can provide as good performances as traditional batch learning techniques, but incorporating the advantages of incremental learning for designing complex pattern recognition systems. Experiments on handwritten characters have shown that a proposed variant of the ensemble training algorithm, employing ensembles of HMMs, can lead to very promising performances. Furthermore, the use of a validation dataset demonstrated that it is possible to reach better performances than the ones presented by batch learning.     

环状物新检测方法及在敏感图像识别中的应用

为了克服传统的Hough变换类环状物体检测的局限,提出了1种结合物体形状与外观特征的环状物体识别检测算法.识别算法使用cascade结构,分别使用灰度、纹理以及外观综合特征,按Bagging的方法训练产生一组弱分类器.这些弱分类器串接而成,并结合局部物体分割,逐个处理当前扫描窗口.相比于传统的Hough算法,新方法具有更快的检测速度.选择敏感图像作为实验对象,采集数据进行训练和检测,实验结果表明,新方法具有明显的性能优势.使用更全面的物体外观信息,按Bagging产生弱分类器的组合,能够在提高环状物体的检测性能的同时,获得理想的处理速度.     

面向目标的带先验概率的AdaBoost算法

针对集成学习算法研究中多个分类器的最佳组合问题,改进了传统的AdaBoost集成学习算法.用于组合的各个分类器通常是基于样本集通过一定的训练得到,样本集中不同类目标的比率可以反映分类目标的先验概率.使用该参数给出了新的组合参数和投票表决阈值计算公式,巧妙的利用样本权值并将其加入到样本属性上进行训练学习,采用新的策略来选择基分类器,给出了面向目标的带先验概率的AdaBoost算法(GWPP AdaBoost算法)和分类器的最佳组合.依据UCI实验数据对传统的AdaBoost算法、Bagging算法、GWPP AdaBoost算法的错误率和性能进行了比较分析,验证了GWPP AdaBoost的有效性.     

Generalized Needleman–Wunsch algorithm for the recognition of T-cell epitopes

In this paper, we propose a new encoding technique that combines the different physicochemical properties of amino acids together with Needleman–Wunsch algorithm. The algorithm was tested in the recognition of T-cell epitopes. A series of SVM classifiers, where each SVM is trained using a different physicochemical property, combined with the “max rule” enables us to obtain an improvement over the state-of-the-art approaches.     

Robust Learning with Missing Data

This paper introduces a new method, called the robust Bayesian estimator (RBE), to learn conditional probability distributions from incomplete data sets. The intuition behind the RBE is that, when no information about the pattern of missing data is available, an incomplete database constrains the set of all possible estimates and this paper provides a characterization of these constraints. An experimental comparison with two popular methods to estimate conditional probability distributions from incomplete data—Gibbs sampling and the EM algorithm—shows a gain in robustness. An application of the RBE to quantify a naive Bayesian classifier from an incomplete data set illustrates its practical relevance.     

多分类器融合的指纹全局特征协同识别

指纹识别是生物特征识别中的热点,指纹全局特征识别具有明显的优势,但是单一分类器一般不能取得满意的识别效果。本文采用贝叶斯理论分析了常见的积、和、中值以及投票多分类器融合方法,并根据实际的选举情形,对投票法进行了2种改进。然后对3种指纹全局特征协同识别分类器:灰度值、主分量以及方向场分类器进行决策层融合,并采用了一种崭新而高效的协同模式识别方法。对FVC2002指纹库的实验表明:该方法具有较好的分类性能,预处理与特征提取简单、计算复杂度低、识别速度快、对污损指纹具有可靠的识别率、鲁棒性强,而且应用于身份认证中也取得了较好的认证效果。     

基于双尺度柔性原型迁移网络的空间滚动轴承寿命阶段识别

针对变工况条件下样本分布差异较大、不同寿命阶段样本数量不均衡导致现有空间滚动轴承寿命阶段识别方法的寿命阶段识别精度较低问题,提出基于双尺度柔性原型迁移网络（Dual scale flexible prototype transfer network,DSFPTN）的空间滚动轴承寿命阶段识别方法。在所提出的DSFPTN中,构造双尺度柔性域感知模块并将其嵌入特征提取器来增强特征提取器对不同领域私有特征的探索能力,提高特征提取器对空间滚动轴承源域和目标域样本特征的学习能力;设计同域泛原型学习以防止跨域样本不加区分的特征学习和不正确聚类,增加两域异类样本的区分性;构建两域原型迁移机制来获得域不变原型,实现从源域原型到目标域原型的迁移;利用加载域不变原型后的双分类器对齐两域之间的分布并计算目标域待测样本与域不变原型之间相似度完成对空间滚动轴承目标域待测样本分类,该分类方式在不同寿命阶段样本数量不均衡条件下能提高对各寿命阶段样本的识别精度。地面模拟空间环境下空间滚动轴承寿命阶段识别实例验证所提出的基于DSFPTN的寿命阶段识别方法的有效性。总之,构建双尺度柔性域感知模块、同域泛原型、两域原型迁移机制和加载域不变原型的双分类器使得DSFPTN在样本分布差异较大以及不同寿命阶段样本数量不均衡条件下,仅利用空间滚动轴承源域的非均衡有标签样本就能对目标域待测样本进行较高精度的寿命阶段识别。