基于重采样和集成选择的适用于实体识别的多分类器系统

实体识别常利用分类器根据记录对的字段相似度向量将记录对分为匹配、不匹配和可能匹配,因此分类器的准确性与实体识别的准确性直接相关。为提高分类准确性,本文基于重采样和集成选择技术构建一个多分类器系统。充分利用实体识别的特点,在分类之前发现分类困难的样本,并使重采样比率在一个区间内变化,生成一组重采样样本;然后用重采样后的样本训练分类器构建一个并行多分类器系统,强调分类器之间的差异度和稀疏度,从该多分类器系统中选择最优分类器子集,即最优的重采样比率组合,分别用非线性规划和极值方法求解该集成选择模型。实验结果表明,本方法与现有的多分类器系统相比具有更高的准确性。     

动态加权投票的多分类器聚合

在使用多分类器系统时,一种流行的方法是采用简单的多数投票策略来聚合多分类器。然而,当各个独立的分类器的性能不统一时,这种简单的多数投票规则会对分类结果造成负面影响。引入一种新的动态加权函数来聚合多个分类器,动态加权函数通过增加分类结果距离样本最近的分类器的权值来提高分类器的性能。在UCI机器学习数据库中的几个现实问题数据集上的实验结果表明动态加权的多分类器聚合方法比简单的多数投票方法能取得更好的分类结果。     

加强概率推理模型在人脸识别中的应用

在很多应用中,组合使用多个分类器可以降低分类错误率。该文就是基于这个思想提出了新的人脸识别算法,即加强概率推理模型。在该算法中,将分类任务划分成多个子分类器,每个子分类器集中于一些难分类的样本,然后组合这些子分类器形成一个强的分类器。试验结果表明算法的识别率比原来的概率推理模型的识别率提高了1.8％。     

基于全信息相关度的动态多分类器融合

AdaB00st采用级联方法生成各基分类器,较好地体现了分类器之间的差异性和互补性.其存在的问题是,在迭代的后期,训练分类器越来越集中在某一小区域的样本上,生成的基分类器体现不同区域的分类特征.根据基分类器的全局分类性能得到固定的投票权重,不能体现基分类器在不同区域上的局部性能差别.因此,本文基于Ada-Boost融合方法,利用待测样本与各分类器的全信息相关度描述基分类器的局部分类性能,提出基于全信息相关度的动态多分类器融合方法,根据各分类器对待测样本的局部分类性能动态确定分类器组合和权重.仿真实验结果表明,该算法提高了融合分类性能.     

基于专家域的多层分类器融合

论文提出了一种基于专家域的多层分类器融合模型,专家指不同专长之单分类器。模型思想来自医院诊断流程,模型首先训练n个专家,之后将样本空间按专家专长划分专家域。对于待测样本,先将样本指派到合适的专家域,然后再由指定的专家对样本进行分类。用这种算法对UCI的标准数据集进行分类,实验结果显示,该算法得到比其他算法更低的分类误差,显著提高了分类器的性能。     

基于多分类器协同学习的卷积神经网络训练算法

卷积神经网络(CNN)是一类重要的深度神经网络,然而其训练过程需要大量的已标记样本,从而限制了其实际应用。针对这一问题,分析了CNN分类器的协同学习过程,给出了基于迭代进化的分类器协同训练算法CAMC。该算法结合了CNN和多分类器协同训练的优势,首先采用不同的卷积核提取出多种样本特征以产生不同的CNN分类器;然后利用少量的已标记样本和大量的未标记样本对多个分类器进行协同训练,以持续提高分类性能。在人脸表情标准数据集上的实验结果表明,相对于传统的表情特征识别法LBP和Gabor,CAMC能够在分类过程中利用未标记样本持续实现性能提升,从而具有更高的分类准确率。     

样本自适应多特征加权的高分辨率遥感图像分类

高分辨率遥感影像能够提供丰富的地物细节,但各种地物空间分布复杂,同类目标呈现出较大的光谱异质性,给传统模式识别分类器带来极大的挑战。提出了一种样本自适应多特征加权的遥感图像分类方法。常见的多特征组合分类器未能充分利用各种特征之间的局部相关性,提出通过分析测试样本局部特征相关性,探究各个特征在不同样本的分类中所占权重的不同,据此对不同分类器进行自适应加权。在一个大型遥感图像数据库上的实验结果表明,不同特征在遥感图像中对不同样本的分类作用是不同的,样本自适应特征加权法将平均分类精度从78.3%提高到90%。     

一种基于预分类的高效最近邻分类器算法

本文的最近邻分类器算法是采用多分类器组合的方式对测试样本进行预分类，并根据预分类结果重新生成新的训练和测试样本集。对新的测试样本采用最近邻分类器进行分类识别，并将识别结果与预分类结果结合在一起进行正确率测试。在ORL人脸库上的实验结果说明，该算法对小样本数据的识别具有明显优势。     

基于BSMOTE和逆转欠抽样的不均衡数据分类算法

针对传统分类器在数据不均衡的情况下分类效果不理想的缺陷,为提高分类器在不均衡数据集下的分类性能,特别是少数类样本的分类能力,提出了一种基于BSMOTE 和逆转欠抽样的不均衡数据分类算法。该算法使用BSMOTE进行过抽样,人工增加少数类样本的数量,然后通过优先去除样本中的冗余和噪声样本,使用逆转欠抽样方法逆转少数类样本和多数类样本的比例。通过多次进行上述抽样形成多个训练集合,使用Bagging方法集成在多个训练集合上获得的分类器来提高有效信息的利用率。实验表明,该算法较几种现有算法不仅能够提高少数类样本的分类性能,而且能够有效提高整体分类准确度。     

多类型分类器融合的文本分类方法研究

传统的文本分类方法大多数使用单一的分类器,而不同的分类器对分类任务的侧重点不同,就使得单一的分类方法有一定的局限性,同时每个特征提取方法对特征词的考虑角度不同。针对以上问题,提出了多类型分类器融合的文本分类方法。该模型使用了word2vec、主成分分析、潜在语义索引以及TFIDF特征提取方法作为多类型分类器融合的特征提取方法。并在多类型分类器加权投票方法中忽略了类别信息的问题,提出了类别加权的分类器权重计算方法。通过实验结果表明,多类型分类器融合方法在二元语料库、多元语料库以及特定语料库上都取得了很好的性能,类别加权的分类器权重计算方法比多类型分类器融合方法在分类性能方面提高了1.19%。     

一种有监督学习证据理论分类器

经典的证据理论不包括从实例中学习基本信度分配的机制,因此应用范围受到一定限制。通过在证据理论中引入神经网络的学习机制,该文提出了一种有监督学习证据理论分类器。该分类器使用一种经过修改的Widrow-Hoff学习规则从训练实例中学习基本信度分配信息。新实例到来后,该分类器在所学基本信度分配的基础上,使用证据理论合成公式对新实例作分类。新分类器拓展了证据理论的应用领域。实验结果表明该分类器是有效的。     

Classifier combination based on confidence transformation

This paper investigates the effects of confidence transformation in combining multiple classifiers using various combination rules. The combination methods were tested in handwritten digit recognition by combining varying classifier sets. The classifier outputs are transformed to confidence measures by combining three scaling functions (global normalization, Gaussian density modeling, and logistic regression) and three confidence types (linear, sigmoid, and evidence). The combination rules include fixed rules (sum-rule, product-rule, median-rule, etc.) and trained rules (linear discriminants and weighted combination with various parameter estimation techniques). The experimental results justify that confidence transformation benefits the combination performance of either fixed rules or trained rules. Trained rules mostly outperform fixed rules, especially when the classifier set contains weak classifiers. Among the trained rules, the support vector machine with linear kernel (linear SVM) performs best while the weighted combination with optimized weights performs comparably well. I have also attempted the joint optimization of confidence parameters and combination weights but its performance was inferior to that of cascaded confidence transformation-combination. This justifies that the cascaded strategy is a right way of multiple classifier combination.     

Evolving decision tree rule based system for audio stego anomalies detection based on Hausdorff distance statistics

This work is motivated by the interest in forensics steganalysis which is aimed at detecting the presence of secret messages transmitted through a subliminal channel. A critical part of the steganalyser design depends on the choice of stego-sensitive features and an efficient machine learning paradigm. The goals of this paper are: (1) to demonstrate that the higher-order statistics of Hausdorff distance - a dissimilarity metric, offers potential discrimination ability for a clean and a stego audio and (2) to achieve promising classification accuracy by realizing the proposed steganalyser with evolving decision tree classifier. Stego sensitive feature selection process is imparted by the genetic algorithm (GA) component and the construction of the rule base is facilitated by the decision tree module. The objective function is designed to maximize the Precision and Recall measures of the classifier thereby enhancing the detection accuracy of the system with low-dimensional and informative features. An extensive experimental evaluation of the proposed system on a database containing 4800 clean and stego audio files (generated by using six different embedding schemes), with the family of six GA decision trees was conducted. The observations reported as 90%+ detection rate, a promising score for a blind steganalyser, show that the proposed scheme, with the Hausdorff distance statistics as features and the evolving decision tree as classifier, is a state-of-the-art steganalyser that outperforms many of the previous steganalytic methods.     

The combination of multiple classifiers using an evidential reasoning approach

In many domains when we have several competing classifiers available we want to synthesize them or some of them to get a more accurate classifier by a combination function. In this paper we propose a ‘class-indifferent’ method for combining classifier decisions represented by evidential structures called triplet and quartet, using Dempster's rule of combination. This method is unique in that it distinguishes important elements from the trivial ones in representing classifier decisions, makes use of more information than others in calculating the support for class labels and provides a practical way to apply the theoretically appealing Dempster-Shafer theory of evidence to the problem of ensemble learning. We present a formalism for modelling classifier decisions as triplet mass functions and we establish a range of formulae for combining these mass functions in order to arrive at a consensus decision. In addition we carry out a comparative study with the alternatives of simplet and dichotomous structure and also compare two combination methods, Dempster's rule and majority voting, over the UCI benchmark data, to demonstrate the advantage our approach offers.     

基于主分量分类的交通事件自动检测算法

利用主分量分类方法,研究改进的基于主分量分类的交通事件自动检测算法。主分量分类方法是一种改进的两类模型分类法。该分类法求解样本方向,该方向可以看作超平面的法方向,根据这个方向将样本中一类数据从另一类数据中分离。样本在法方向上的投影用来估计每个实例的条件概率,然后根据贝叶斯规则实现实例的分类。对于线性不可分等复杂的分类问题,可通过核函数作用将数据映射到高维特征空间中实现线性可分。最后对I-880高速公路事件数据的仿真结果表明,KPCC算法获得了100.00%的检测率、1.82%的误警率和1.02分钟的平均检测时间。     

On the independence requirement in Dempster-Shafer theory for combining classifiers providing statistical evidence

In classifier combination, the relative values of beliefs assigned to different hypotheses are more important than accurate estimation of the combined belief function representing the joint observation space. Because of this, the independence requirement in Dempster’s rule should be examined from classifier combination point of view. In this study, it is investigated whether there is a set of dependent classifiers which provides a better combined accuracy than independent classifiers when Dempster’s rule of combination is used. The analysis carried out for three different representations of statistical evidence has shown that the combination of dependent classifiers using Dempster’s rule may provide much better combined accuracies compared to independent classifiers.     

A Component-based Framework for Face Detection and Identification

We present a component-based framework for face detection and identification. The face detection and identification modules share the same hierarchical architecture. They both consist of two layers of classifiers, a layer with a set of component classifiers and a layer with a single combination classifier. The component classifiers independently detect/identify facial parts in the image. Their outputs are passed the combination classifier which performs the final detection/identification of the face. We describe an algorithm which automatically learns two separate sets of facial components for the detection and identification tasks. In experiments we compare the detection and identification systems to standard global approaches. The experimental results clearly show that our component-based approach is superior to global approaches.     

An experimental study of one- and two-level classifier fusion for different sample sizes

Due to the wide variety of fusion techniques available for combining multiple classifiers into a more accurate classifier, a number of good studies have been devoted to determining in what situations some fusion methods should be preferred over other ones. However, the sample size behavior of the various fusion methods has hitherto received little attention in the literature of multiple classifier systems. The main contribution of this paper is thus to investigate the effect of training sample size on their relative performance and to gain more insight into the conditions for the superiority of some combination rules.A large experiment is conducted to study the performance of some fixed and trainable combination rules for executing one- and two-level classifier fusion for different training sample sizes. The experimental results yield the following conclusions: when implementing one-level fusion to combine homogeneous or heterogeneous base classifiers, fixed rules outperform trainable ones in nearly all cases, with only one exception of merging heterogeneous classifiers for large sample size. Moreover, the best classification for any considered sample size is generally achieved by a second level of combination (namely, utilizing one fusion rule to further combine a set of ensemble classifiers with each of them constructed by fusing base classifiers). Under these circumstances, it seems that adopting different types of fusion rules (fixed or trainable) as the combiners for two levels of fusion is appropriate.     

New Applications of Ensembles of Classifiers

Combination (ensembles) of classifiers is now a well established research line. It has been observed that the predictive accuracy of a combination of independent classifiers excels that of the single best classifier. While ensembles of classifiers have been mostly employed to achieve higher recognition accuracy, this paper focuses on the use of combinations of individual classifiers for handling several problems from the practice in the machine learning, pattern recognition and data mining domains. In particular, the study presented concentrates on managing the imbalanced training sample problem, scaling up some preprocessing algorithms and filtering the training set. Here, all these situations are examined mainly in connection with the nearest neighbour classifier. Experimental results show the potential of multiple classifier systems when applied to those situations.     

基于OAI-PMH架构的元数据分类研究

提出了一种基于规则和学习相结合的元数据分类存储的方法，该方法通过统计分析，提取对元数据分类影响较大的通用特征规则，对无法用规则分类的元数据，采用文本学习方法，将元数据记录看成由多个属性字段组成的文本，通过构造分类器实现分类。实验结果表明，采用元数据分类存储方法具有良好的检索性能。