Decision trees for hierarchical multi-label classification

Hierarchical multi-label classification (HMC) is a variant of classification where instances may belong to multiple classes at the same time and these classes are organized in a hierarchy. This article presents several approaches to the induction of decision trees for HMC, as well as an empirical study of their use in functional genomics. We compare learning a single HMC tree (which makes predictions for all classes together) to two approaches that learn a set of regular classification trees (one for each class). The first approach defines an independent single-label classification task for each class (SC). Obviously, the hierarchy introduces dependencies between the classes. While they are ignored by the first approach, they are exploited by the second approach, named hierarchical single-label classification (HSC). Depending on the application at hand, the hierarchy of classes can be such that each class has at most one parent (tree structure) or such that classes may have multiple parents (DAG structure). The latter case has not been considered before and we show how the HMC and HSC approaches can be modified to support this setting. We compare the three approaches on 24 yeast data sets using as classification schemes MIPS’s FunCat (tree structure) and the Gene Ontology (DAG structure). We show that HMC trees outperform HSC and SC trees along three dimensions: predictive accuracy, model size, and induction time. We conclude that HMC trees should definitely be considered in HMC tasks where interpretable models are desired.     

Applying semi-supervised learning in hierarchical multi-label classification

In classification problems with hierarchical structures of labels, the target function must assign labels that are hierarchically organized and it can be used either for single-label (one label per instance) or multi-label classification problems (more than one label per instance). In parallel to these developments, the idea of semi-supervised learning has emerged as a solution to the problems found in a standard supervised learning procedure (used in most classification algorithms). It combines labelled and unlabelled data during the training phase. Some semi-supervised methods have been proposed for single-label classification methods. However, very little effort has been done in the context of multi-label hierarchical classification. Therefore, this paper proposes a new method for supervised hierarchical multi-label classification, called HMC-RAkEL. Additionally, we propose the use of semi-supervised learning, self-training, in hierarchical multi-label classification, leading to three new methods, called HMC-SSBR, HMC-SSLP and HMC-SSRAkEL. In order to validate the feasibility of these methods, an empirical analysis will be conducted, comparing the proposed methods with their corresponding supervised versions. The main aim of this analysis is to observe whether the semi-supervised methods proposed in this paper have similar performance of the corresponding supervised versions.     

Incorporating label dependency into the binary relevance framework for multi-label classification

In multi-label classification, examples can be associated with multiple labels simultaneously. The task of learning from multi-label data can be addressed by methods that transform the multi-label classification problem into several single-label classification problems. The binary relevance approach is one of these methods, where the multi-label learning task is decomposed into several independent binary classification problems, one for each label in the set of labels, and the final labels for each example are determined by aggregating the predictions from all binary classifiers. However, this approach fails to consider any dependency among the labels. Aiming to accurately predict label combinations, in this paper we propose a simple approach that enables the binary classifiers to discover existing label dependency by themselves. An experimental study using decision trees, a kernel method as well as Naïve Bayes as base-learning techniques shows the potential of the proposed approach to improve the multi-label classification performance.     

Enhancing multi-label classification by modeling dependencies among labels

In this paper, we propose a novel framework for multi-label classification, which directly models the dependencies among labels using a Bayesian network. Each node of the Bayesian network represents a label, and the links and conditional probabilities capture the probabilistic dependencies among multiple labels. We employ our Bayesian network structure learning method, which guarantees to find the global optimum structure, independent of the initial structure. After structure learning, maximum likelihood estimation is used to learn the conditional probabilities among nodes. Any current multi-label classifier can be employed to obtain the measurements of labels. Then, using the learned Bayesian network, the true labels are inferred by combining the relationship among labels with the labels? estimates obtained from a current multi-labeling method. We further extend the proposed multi-label classification method to deal with incomplete label assignments. Structural Expectation-Maximization algorithm is adopted for both structure and parameter learning. Experimental results on two benchmark multi-label databases show that our approach can effectively capture the co-occurrent and the mutual exclusive relation among labels. The relation modeled by our approach is more flexible than the pairwise or fixed subset labels captured by current multi-label learning methods. Thus, our approach improves the performance over current multi-label classifiers. Furthermore, our approach demonstrates its robustness to incomplete multi-label classification.     

基于堆叠模型的司法短文本多标签分类

司法文书短文本的语义多样性和特征稀疏性等特点,对短文本多标签分类精度提出了很大的挑战,传统单一模型的分类算法已无法满足业务需求。为此,提出一种融合深度学习与堆叠模型的多标签分类方法。该方法将分类器划分成两个层次,第一层使用BERT、卷积神经网络、门限循环单元等深度学习方法作为基础分类器,每个基础分类器模型通过K折交叉验证得到所有数据的多标签分类概率值,将此概率值数据进行融合形成元数据;第二层使用自定义的深度神经网络作为混合器,以第一层的元数据为输入,通过训练多标签概率矩阵获取模型参数。该方法将强分类器关联在一起,获得比单个分类器更加强大的性能。实验结果表明,深度学习堆叠模型实现了87%左右的短文本分类F1分数,优于BERT、卷积神经网络、循环神经网络及其他单个模型的性能。     

Hierarchical classification of protein function with ensembles of rules and particle swarm optimisation

This paper focuses on hierarchical classification problems where the classes to be predicted are organized in the form of a tree. The standard top-down divide and conquer approach for hierarchical classification consists of building a hierarchy of classifiers where a classifier is built for each internal (non-leaf) node in the class tree. Each classifier discriminates only between its child classes. After the tree of classifiers is built, the system uses them to classify test examples one class level at a time, so that when the example is assigned a class at a given level, only the child classes need to be considered at the next level. This approach has the drawback that, if a test example is misclassified at a certain class level, it will be misclassified at deeper levels too. In this paper we propose hierarchical classification methods to mitigate this drawback. More precisely, we propose a method called hierarchical ensemble of hierarchical rule sets (HEHRS), where different ensembles are built at different levels in the class tree and each ensemble consists of different rule sets built from training examples at different levels of the class tree. We also use a particle swarm optimisation (PSO) algorithm to optimise the rule weights used by HEHRS to combine the predictions of different rules into a class to be assigned to a given test example. In addition, we propose a variant of a method to mitigate the aforementioned drawback of top-down classification. These three types of methods are compared against the standard top-down hierarchical classification method in six challenging bioinformatics datasets, involving the prediction of protein function. Overall HEHRS with the rule weights optimised by the PSO algorithm obtains the best predictive accuracy out of the four types of hierarchical classification method.     

An efficient multi-label support vector machine with a zero label

Existing multi-label support vector machine (Rank-SVM) has an extremely high computational complexity and lacks an intrinsic zero point to determine relevant labels. In this paper, we propose a novel support vector machine for multi-label classification through both simplifying Rank-SVM and adding a zero label, resulting into a quadratic programming problem in which each class has an independent equality constraint. When Frank-Wolfe method is used to solve our quadratic programming problem iteratively, our entire linear programming problem of each step is divided into a series of sub-problems, which dramatically reduces computational cost. It is illustrated that for famous Yeast data set our training procedure runs about 12 times faster than Rank-SVM does under C++ environment. Experiments from five benchmark data sets show that our method is a powerful candidate for multi-label classification, compared with five state-of-the-art multi-label classification techniques.     

Multi layer ELM-RBF for multi-label learning

Many neural network methods such as ML-RBF and BP-MLL have been used for multi-label classification. Recently, extreme learning machine (ELM) is used as the basic elements to handle multi-label classification problem because of its fast training time. Extreme learning machine based auto encoder (ELM-AE) is a novel method of neural network which can reproduce the input signal as well as auto encoder, but it can not solve the over-fitting problem in neural networks elegantly. Introducing weight uncertainty into ELM-AE, we can treat the input weights as random variables following Gaussian distribution and propose weight uncertainty ELM-AE (WuELM-AE). In this paper, a neural network named multi layer ELM-RBF for multi-label learning (ML-ELM-RBF) is proposed. It is derived from radial basis function for multi-label learning (ML-RBF) and WuELM-AE. ML-ELM-RBF firstly stacks WuELM-AE to create a deep network, and then it conducts clustering analysis on samples features of each possible class to compose the last hidden layer. ML-ELM-RBF has achieved satisfactory results on single-label and multi-label data sets. Experimental results show that WuELM-AE and ML-ELM-RBF are effective learning algorithms.     

基于注意力与标签相关性的胸部X光片疾病分类

针对传统的胸部辅助诊断系统在胸部X光片疾病分类方面图像特征提取效果差、平均准确率低等问题,提出了一个注意力机制和标签相关性结合的多层次分类网络.网络的训练分为两个阶段,在阶段1为了提高网络特征提取能力,引入注意力机制并构建一个双分支特征提取网络,实现综合特征的提取,在阶段2考虑到多标签分类中标签之间相关性等问题,利用图卷积神经网络对标签相关关系进行建模,并与阶段1的特征提取结果进行结合,以实现对胸部X光片疾病的多标签分类任务.实验结果表明,本方法在ChestX-ray14数据集上各类疾病的加权平均AUC达到0.827,有助于辅助医生进行胸部疾病的诊断,有一定的临床应用价值.     

基于多标签神经网络的行人属性识别

在多标签行人属性识别的问题中,为了充分利用标签之间的相关性,解决传统方法识别准确率低和效率慢的问题,提出了一个多标签卷积神经网络。该网络在一个统一的网络框架下识别行人多个属性。把行人的多个属性看作是一个序列,然后构建了一个时序分类模型。提出的方法不仅避免了复杂的多输入MLCNN网络,也不需要多次训练单标签分类模型。实验结果表明,本文方法准确率均优于SIFT+SVM和多输入的MLCNN模型,平均准确率达到了90.41%。     

Predicting adequacy of vancomycin regimens: A learning-based classification approach to improving clinical decision making

Clinicians' drug regimen decision making is critical, particularly when involving high-alert medications. In this study, we use decision-tree induction C4.5 and a backpropagation neural network to construct decision support systems for predicting the regimen adequacy of vancomycin, a glycopeptide antimicrobial antibiotic effective for Gram-positive bacterial infections. We comparatively evaluate the respective systems using a total of 987 clinical vancomycin cases collected from a major tertiary medical center in southern Taiwan. We supplement each system using Bagging and then examine the predictive power of the extended system. Overall, our evaluation results show the overall accuracy of the decision support system based on C4.5 or the neural network to be significantly higher than that of the benchmark one-compartment pharmacokinetic model. Use of Bagging can considerably improve the effectiveness of each system across different performance measures, particularly for cases of decision classes in which the base systems (i.e., without Bagging) perform modestly. Taken together, our evaluation results seem to favor the use of Bagging to enhance the performance of decision support systems constructed using decision-tree induction C4.5.     

An extended one-versus-rest support vector machine for multi-label classification

Hybrid strategy, which generalizes a specific single-label algorithm while one or two data decomposition tricks are applied implicitly or explicitly, has become an effective and efficient tool to design and implement various multi-label classification algorithms. In this paper, we extend traditional binary support vector machine by introducing an approximate ranking loss as its empirical loss term to build a novel support vector machine for multi-label classification, resulting into a quadratic programming problem with different upper bounds of variables to characterize label correlation of individual instance. Further, our optimization problem can be solved via combining one-versus-rest data decomposition trick with modified binary support vector machine, which dramatically reduces computational cost. Experimental study on ten multi-label data sets illustrates that our method is a powerful candidate for multi-label classification, compared with four state-of-the-art multi-label classification approaches.     

A hybrid algorithm for Bayesian network structure learning with application to multi-label learning

We present a novel hybrid algorithm for Bayesian network structure learning, called H2PC. It first reconstructs the skeleton of a Bayesian network and then performs a Bayesian-scoring greedy hill-climbing search to orient the edges. The algorithm is based on divide-and-conquer constraint-based subroutines to learn the local structure around a target variable. We conduct two series of experimental comparisons of H2PC against Max–Min Hill-Climbing (MMHC), which is currently the most powerful state-of-the-art algorithm for Bayesian network structure learning. First, we use eight well-known Bayesian network benchmarks with various data sizes to assess the quality of the learned structure returned by the algorithms. Our extensive experiments show that H2PC outperforms MMHC in terms of goodness of fit to new data and quality of the network structure with respect to the true dependence structure of the data. Second, we investigate H2PC’s ability to solve the multi-label learning problem. We provide theoretical results to characterize and identify graphically the so-called minimal label powersets that appear as irreducible factors in the joint distribution under the faithfulness condition. The multi-label learning problem is then decomposed into a series of multi-class classification problems, where each multi-class variable encodes a label powerset. H2PC is shown to compare favorably to MMHC in terms of global classification accuracy over ten multi-label data sets covering different application domains. Overall, our experiments support the conclusions that local structural learning with H2PC in the form of local neighborhood induction is a theoretically well-motivated and empirically effective learning framework that is well suited to multi-label learning. The source code (in R) of H2PC as well as all data sets used for the empirical tests are publicly available.     

Multi-label classification and extracting predicted class hierarchies

This paper investigates hierarchy extraction from results of multi-label classification (MC). MC deals with instances labeled by multiple classes rather than just one, and the classes are often hierarchically organized. Usually multi-label classifiers rely on a predefined class hierarchy. A much less investigated approach is to suppose that the hierarchy is unknown and to infer it automatically. In this setting, the proposed system classifies multi-label data and extracts a class hierarchy from multi-label predictions. It is based on a combination of a novel multi-label extension of the fuzzy Adaptive Resonance Associative Map (ARAM) neural network with an association rule learner.     

基于XDense-RC-net网络的CXR图像分类算法

卷积神经网络逐渐应用于胸部X射线（chset X-ray,CXR）图像分类领域,目前普遍使用迁移学习技术进行分类研究,快速构建网络的同时未能考虑CXR图像的特异性。针对上述问题,提出了一种新型的XDense-RC-net方法。该方法对DenseNet模型进行改进,在原密集连接层引入新提出的空间注意力机制,实现特征提取和特征融合,优化DenseNet的transition模块,同时使用两种不同的池化策略增强模型的抗扰动能力。实验使用chest X-ray14多标签14分类数据集和COVIDx单标签3分类数据集对XDense-RC-net进行验证。在多标签分类实验中,平均AUC值达到0.854,比基准方法提升了0.109。在单标签分类实验中,平均准确率达到96.75%,相较于基准方法提升了7.75%。结果显示,XDense-RC-net提升了CXR图像分类的精度,并能够泛化至多标签和单标签两种不同的分类任务中。     

一种循环卷积注意力模型的文本情感分类方法

情感分类对推荐系统、自动问答、阅读理解等下游应用具有重要应用价值,是自然语言处理领域的重要研究方向。情感分类任务直接依赖于上下文,包括全局和局部信息,而现有的神经网络模型无法同时捕获上下文局部信息和全局信息。文中针对单标记和多标记情感分类任务,提出一种循环卷积注意力模型(LSTM-CNN-ATT,LCA)。该模型利用注意力机制融合卷积神经网络(Convolutional Neural Network,CNN)的局部信息提取能力和循环神经网络(Recurrent Neural Network,RNN)的全局信息提取能力,包括词嵌入层、上下文表示层、卷积层和注意力层。对于多标记情感分类任务,在注意力层上附加主题信息,进一步指导多标记情感倾向的精确提取。在两个单标记数据集上的F1指标达到82.1%,与前沿单标记模型相当;在两个多标记数据集上,小数据集实验结果接近基准模型,大数据集上的F1指标达到78.38%,超过前沿模型,表明LCA模型具有较高的稳定性和较强的通用性。     

多标签文本分类研究进展

文本分类作为自然语言处理中一个基本任务,在20世纪50年代就已经对其算法进行了研究,现在单标签文本分类算法已经趋向成熟,但是对于多标签文本分类的研究还有很大的提升空间。介绍了多标签文本分类的基本概念以及基本流程,包括数据集获取、文本预处理、模型训练和预测结果。介绍了多标签文本分类的方法。这些方法主要分为两大类：传统机器学习方法和基于深度学习的方法。传统机器学习方法主要包括问题转换方法和算法自适应方法。基于深度学习的方法是利用各种神经网络模型来处理多标签文本分类问题,根据模型结构,将其分为基于CNN结构、基于RNN结构和基于Transfomer结构的多标签文本分类方法。对多标签文本分类常用的数据集进行了梳理总结。对未来的发展趋势进行了分析与展望。     

基于改进卷积神经网络的多标记分类算法

良好的特征表达是提高模型性能的关键，然而当前在多标记学习领域，特征表达依然采用人工设计的方式，所提取的特征抽象程度不高，包含的可区分性信息不足。针对此问题，提出了基于卷积神经网络的多标记分类模型ML_DCCNN，该模型利用卷积神经网络强大的特征提取能力，自动学习能刻画数据本质的特征。为了解决深度卷积神经网络预测精度高，但训练时间复杂度不低的问题，ML_DCCNN利用迁移学习方法缩减模型的训练时间，同时改进卷积神经网络的全连接层，提出双通道神经元，减少全连接层的参数量。实验表明，与传统的多标记分类算法以及已有的基于深度学习的多标记分类模型相比，ML_DCCNN保持了较高的分类精度并有效地提高了分类效率，具有一定的理论与实际价值。