Cost-sensitive learning with conditional Markov networks

There has been a recent, growing interest in classification and link prediction in structured domains. Methods such as conditional random fields and relational Markov networks support flexible mechanisms for modeling correlations due to the link structure. In addition, in many structured domains, there is an interesting structure in the risk or cost function associated with different misclassifications. There is a rich tradition of cost-sensitive learning applied to unstructured (IID) data. Here we propose a general framework which can capture correlations in the link structure and handle structured cost functions. We present two new cost-sensitive structured classifiers based on maximum entropy principles. The first determines the cost-sensitive classification by minimizing the expected cost of misclassification. The second directly determines the cost-sensitive classification without going through a probability estimation step. We contrast these approaches with an approach which employs a standard 0/1-loss structured classifier to estimate class conditional probabilities followed by minimization of the expected cost of misclassification and with a cost-sensitive IID classifier that does not utilize the correlations present in the link structure. We demonstrate the utility of our cost-sensitive structured classifiers with experiments on both synthetic and real-world data.     

Improving specific class mapping from remotely sensed data by cost-sensitive learning

In many remote-sensing projects, one is usually interested in a small number of land-cover classes present in a study area and not in all the land-cover classes that make-up the landscape. Previous studies in supervised classification of satellite images have tackled specific class mapping problem by isolating the classes of interest and combining all other classes into one large class, usually called others, and by developing a binary classifier to discriminate the class of interest from the others. Here, this approach is called focused approach. The strength of the focused approach is to decompose the original multi-class supervised classification problem into a binary classification problem, focusing the process on the discrimination of the class of interest. Previous studies have shown that this method is able to discriminate more accurately the classes of interest when compared with the standard multi-class supervised approach. However, it may be susceptible to data imbalance problems present in the training data set, since the classes of interest are often a small part of the training set. A result the classification may be biased towards the largest classes and, thus, be sub-optimal for the discrimination of the classes of interest. This study presents a way to minimize the effects of data imbalance problems in specific class mapping using cost-sensitive learning. In this approach errors committed in the minority class are treated as being costlier than errors committed in the majority class. Cost-sensitive approaches are typically implemented by weighting training data points accordingly to their importance to the analysis. By changing the weight of individual data points, it is possible to shift the weight from the larger classes to the smaller ones, balancing the data set. To illustrate the use of the cost-sensitive approach to map specific classes of interest, a series of experiments with weighted support vector machines classifier and Landsat Thematic Mapper data were conducted to discriminate two types of mangrove forest (high-mangrove and low-mangrove) in Saloum estuary, Senegal, a United Nations Educational, Scientific and Cultural Organisation World Heritage site. Results suggest an increase in overall classification accuracy with the use of cost-sensitive method (97.3%) over the standard multi-class (94.3%) and the focused approach (91.0%). In particular, cost-sensitive method yielded higher sensitivity and specificity values on the discrimination of the classes of interest when compared with the standard multi-class and focused approaches.     

一种基于重取样的代价敏感学习算法

大多数非均衡数据集的研究集中于纯重构数据集或者纯代价敏感学习,本文针对数据集类分布非均衡和不相等误分类代价往往同时发生这一事实,提出了一种以最小误分类代价为目标的基于混合重取样的代价敏感学习算法。该算法将两种不同类型解决方案有机地融合在一起,先用样本类空间重构的方法使原始数据集的两类数据达到基本均衡,然后再引入代价敏感学习算法进行分类,能提高少数类分类精度,同时有效降低总的误分类代价。实验结果验证了该算法在处理非均衡类问题时比传统算法要优越。     

Handling over-fitting in test cost-sensitive decision tree learning by feature selection, smoothing and pruning

Cost-sensitive learning algorithms are typically designed for minimizing the total cost when multiple costs are taken into account. Like other learning algorithms, cost-sensitive learning algorithms must face a significant challenge, over-fitting, in an applied context of cost-sensitive learning. Specifically speaking, they can generate good results on training data but normally do not produce an optimal model when applied to unseen data in real world applications. It is called data over-fitting. This paper deals with the issue of data over-fitting by designing three simple and efficient strategies, feature selection, smoothing and threshold pruning, against the TCSDT (test cost-sensitive decision tree) method. The feature selection approach is used to pre-process the data set before applying the TCSDT algorithm. The smoothing and threshold pruning are used in a TCSDT algorithm before calculating the class probability estimate for each decision tree leaf. To evaluate our approaches, we conduct extensive experiments on the selected UCI data sets across different cost ratios, and on a real world data set, KDD-98 with real misclassification cost. The experimental results show that our algorithms outperform both the original TCSDT and other competing algorithms on reducing data over-fitting.     

Cost-sensitive label embedding for multi-label classification

Label embedding (LE) is an important family of multi-label classification algorithms that digest the label information jointly for better performance. Different real-world applications evaluate performance by different cost functions of interest. Current LE algorithms often aim to optimize one specific cost function, but they can suffer from bad performance with respect to other cost functions. In this paper, we resolve the performance issue by proposing a novel cost-sensitive LE algorithm that takes the cost function of interest into account. The proposed algorithm, cost-sensitive label embedding with multidimensional scaling (CLEMS), approximates the cost information with the distances of the embedded vectors by using the classic multidimensional scaling approach for manifold learning. CLEMS is able to deal with both symmetric and asymmetric cost functions, and effectively makes cost-sensitive decisions by nearest-neighbor decoding within the embedded vectors. We derive theoretical results that justify how CLEMS achieves the desired cost-sensitivity. Furthermore, extensive experimental results demonstrate that CLEMS is significantly better than a wide spectrum of existing LE algorithms and state-of-the-art cost-sensitive algorithms across different cost functions.     

基于支持向量机的代价敏感挖掘

针对一些数据挖掘应用中反例样本和正例样本具有不同误分类代价的情况，提出一种代价敏感支持向量机算法CS-SVM．CS-SVM包括3个步骤：首先，引入Sigmoid函数，根据样本到分类超平面的距离估计其后验概率；然后，根据误分类代价最小原则重构训练样本的类标号；最后，在重构后的训练集上使用标准SVM进行学习即得到嵌入误分类代价的最优分类超平面．基于CS-SVM的思路，提出一个通用的嵌入误分类代价的代价敏感分类算法G-CSC．试验结果表明：相比于SVM，CS-SVM大大降低测试集上的平均误分类代价．     

Training cost-sensitive neural networks with methods addressing the class imbalance problem

This paper studies empirically the effect of sampling and threshold-moving in training cost-sensitive neural networks. Both oversampling and undersampling are considered. These techniques modify the distribution of the training data such that the costs of the examples are conveyed explicitly by the appearances of the examples. Threshold-moving tries to move the output threshold toward inexpensive classes such that examples with higher costs become harder to be misclassified. Moreover, hard-ensemble and soft-ensemble, i.e., the combination of above techniques via hard or soft voting schemes, are also tested. Twenty-one UCl data sets with three types of cost matrices and a real-world cost-sensitive data set are used in the empirical study. The results suggest that cost-sensitive learning with multiclass tasks is more difficult than with two-class tasks, and a higher degree of class imbalance may increase the difficulty. It also reveals that almost all the techniques are effective on two-class tasks, while most are ineffective and even may cause negative effect on multiclass tasks. Overall, threshold-moving and soft-ensemble are relatively good choices in training cost-sensitive neural networks. The empirical study also suggests that some methods that have been believed to be effective in addressing the class imbalance problem may, in fact, only be effective on learning with imbalanced two-class data sets.     

基于新型不纯度度量的代价敏感随机森林分类器

针对不平衡数据集的有效分类问题,提出一种结合代价敏感学习和随机森林算法的分类器。首先提出了一种新型不纯度度量,该度量不仅考虑了决策树的总代价,还考虑了同一节点对于不同样本的代价差异;其次,执行随机森林算法,对数据集作K次抽样,构建K个基础分类器;然后,基于提出的不纯度度量,通过分类回归树(CART)算法来构建决策树,从而形成决策树森林;最后,随机森林通过投票机制做出数据分类决策。在UCI数据库上进行实验,与传统随机森林和现有的代价敏感随机森林分类器相比,该分类器在分类精度、AUC面积和Kappa系数这3种性能度量上都具有良好的表现。     

PAKDD’12 best paper: generating balanced classifier-independent training samples from unlabeled data

We consider the problem of generating balanced training samples from an unlabeled data set with an unknown class distribution. While random sampling works well when the data are balanced, it is very ineffective for unbalanced data. Other approaches, such as active learning and cost-sensitive learning, are also suboptimal as they are classifier-dependent and require misclassification costs and labeled samples, respectively. We propose a new strategy for generating training samples, which is independent of the underlying class distribution of the data and the classifier that will be trained using the labeled data. Our methods are iterative and can be seen as variants of active learning, where we use semi-supervised clustering at each iteration to perform biased sampling from the clusters. We provide several strategies to estimate the underlying class distributions in the clusters and to increase the balancedness in the training samples. Experiments with both highly skewed and balanced data from the UCI repository and a private data set show that our algorithm produces much more balanced samples than random sampling or uncertainty sampling. Further, our sampling strategy is substantially more efficient than active learning methods. The experiments also validate that, with more balanced training data, classifiers trained with our samples outperform classifiers trained with random sampling or active learning.     

用于不平衡数据分类的代价敏感超网络算法

传统的超网络模型在处理不平衡数据分类问题时,具有很大的偏向性,正类的识别率远远高于负类。为此,提出了一种代价敏感超网络Boosting集成算法。首先,将代价敏感学习引入超网络模型,提出了代价敏感的超网络模型;同时,为了使算法能够自适应正类的错分代价,采用Boosting算法对代价敏感超网络进行集成。代价敏感超网络能很好地修正传统的超网络在处理不平衡数据分类问题时过分偏向正类的缺陷,提高对负类的分类准确性。实验结果表明,代价敏感超网络Boosting集成算法具有处理不平衡数据分类问题的优势。     

Online classifier adaptation for cost-sensitive learning

In this paper, we propose the problem of online cost-sensitive classifier adaptation and the first algorithm to solve it. We assume that we have a base classifier for a cost-sensitive classification problem, but it is trained with respect to a cost setting different to the desired one. Moreover, we also have some training data samples streaming to the algorithm one by one. The problem is to adapt the given base classifier to the desired cost setting using the steaming training samples online. To solve this problem, we propose to learn a new classifier by adding an adaptation function to the base classifier, and update the adaptation function parameter according to the streaming data samples. Given an input data sample and the cost of misclassifying it, we update the adaptation function parameter by minimizing cost-weighted hinge loss and respecting previous learned parameter simultaneously. The proposed algorithm is compared to both online and off-line cost-sensitive algorithms on two cost-sensitive classification problems, and the experiments show that it not only outperforms them on classification performances, but also requires significantly less running time.

     

基于改进的代价敏感决策树的网络贷款分类

在网络贷款用户数据集中，贷款成功和贷款失败的用户数量存在着严重的不平衡，传统的机器学习算法在解决该类问题时注重整体分类正确率，导致贷款成功用户的预测精度较低。针对此问题，在代价敏感决策树敏感函数的计算中加入类分布，以减弱正负样本数量对误分类代价的影响，构建改进的代价敏感决策树；以该决策树作为基分类器并以分类准确度作为衡量标准选择表现较好的基分类器，将它们与最后阶段生成的分类器集成得到最终的分类器。实验结果表明，与已有的常用于解决此类问题的算法（如MetaCost算法、代价敏感决策树、AdaCost算法等）相比，改进的代价敏感决策树对网络贷款用户分类可以降低总体的误分类错误率，具有更强的泛化能力。     

An extended tuning method for cost-sensitive regression and forecasting

In many real-world regression and forecasting problems, over-prediction and under-prediction errors have different consequences and incur asymmetric costs. Such problems entail the use of cost-sensitive learning, which attempts to minimize the expected misprediction cost, rather than minimize a simple measure such as mean squared error. A method has been proposed recently for tuning a regular regression model post hoc so as to minimize the average misprediction cost under an asymmetric cost structure. In this paper, we build upon that method and propose an extended tuning method for cost-sensitive regression. The previous method becomes a special case of the method we propose. We apply the proposed method to loan charge-off forecasting, a cost-sensitive regression problem that has had a bearing on bank failures over the last few years. Empirical evaluation in the loan charge-off forecasting domain demonstrates that the method we have proposed can further lower the misprediction cost significantly.     

Classification cost: An empirical comparison among traditional classifier, Cost-Sensitive Classifier, and MetaCost

Loan fraud is a critical factor in the insolvency of financial institutions, so companies make an effort to reduce the loss from fraud by building a model for proactive fraud prediction. However, there are still two critical problems to be resolved for the fraud detection: (1) the lack of cost sensitivity between type I error and type II error in most prediction models, and (2) highly skewed distribution of class in the dataset used for fraud detection because of sparse fraud-related data. The objective of this paper is to examine whether classification cost is affected both by the cost-sensitive approach and by skewed distribution of class. To that end, we compare the classification cost incurred by a traditional cost-insensitive classification approach and two cost-sensitive classification approaches, Cost-Sensitive Classifier (CSC) and MetaCost. Experiments were conducted with a credit loan dataset from a major financial institution in Korea, while varying the distribution of class in the dataset and the number of input variables. The experiments showed that the lowest classification cost was incurred when the MetaCost approach was used and when non-fraud data and fraud data were balanced. In addition, the dataset that includes all delinquency variables was shown to be most effective on reducing the classification cost.     

A closed-form reduction of multi-class cost-sensitive learning to weighted multi-class learning

In cost-sensitive learning, misclassification costs can vary for different classes. This paper investigates an approach reducing a multi-class cost-sensitive learning to a standard classification task based on the data space expansion technique developed by Abe et al., which coincides with Elkan's reduction with respect to binary classification tasks. Using this proposed reduction approach, a cost-sensitive learning problem can be solved by considering a standard 0/1 loss classification problem on a new distribution determined by the cost matrix. We also propose a new weighting mechanism to solve the reduced standard classification problem, based on a theorem stating that the empirical loss on independently identically distributed samples from the new distribution is essentially the same as the loss on the expanded weighted training set. Experimental results on several synthetic and benchmark datasets show that our weighting approach is more effective than existing representative approaches for cost-sensitive learning.     

Large margin cost-sensitive learning of conditional random fields

We tackle the structured output classification problem using the Conditional Random Fields (CRFs). Unlike the standard 0/1 loss case, we consider a cost-sensitive learning setting where we are given a non-0/1 misclassification cost matrix at the individual output level. Although the task of cost-sensitive classification has many interesting practical applications that retain domain-specific scales in the output space (e.g., hierarchical or ordinal scale), most CRF learning algorithms are unable to effectively deal with the cost-sensitive scenarios as they merely assume a nominal scale (hence 0/1 loss) in the output space. In this paper, we incorporate the cost-sensitive loss into the large margin learning framework. By large margin learning, the proposed algorithm inherits most benefits from the SVM-like margin-based classifiers, such as the provable generalization error bounds. Moreover, the soft-max approximation employed in our approach yields a convex optimization similar to the standard CRF learning with only slight modification in the potential functions. We also provide the theoretical cost-sensitive generalization error bound. We demonstrate the improved prediction performance of the proposed method over the existing approaches in a diverse set of sequence/image structured prediction problems that often arise in pattern recognition and computer vision domains.     

Software defect prediction using cost-sensitive neural network

The software development life cycle generally includes analysis, design, implementation, test and release phases. The testing phase should be operated effectively in order to release bug-free software to end users. In the last two decades, academicians have taken an increasing interest in the software defect prediction problem, several machine learning techniques have been applied for more robust prediction. A different classification approach for this problem is proposed in this paper. A combination of traditional Artificial Neural Network (ANN) and the novel Artificial Bee Colony (ABC) algorithm are used in this study. Training the neural network is performed by ABC algorithm in order to find optimal weights. The False Positive Rate (FPR) and False Negative Rate (FNR) multiplied by parametric cost coefficients are the optimization task of the ABC algorithm. Software defect data in nature have a class imbalance because of the skewed distribution of defective and non-defective modules, so that conventional error functions of the neural network produce unbalanced FPR and FNR results. The proposed approach was applied to five publicly available datasets from the NASA Metrics Data Program repository. Accuracy, probability of detection, probability of false alarm, balance, Area Under Curve (AUC), and Normalized Expected Cost of Misclassification (NECM) are the main performance indicators of our classification approach. In order to prevent random results, the dataset was shuffled and the algorithm was executed 10 times with the use of n-fold cross-validation in each iteration. Our experimental results showed that a cost-sensitive neural network can be created successfully by using the ABC optimization algorithm for the purpose of software defect prediction.     

A <Emphasis Type="Italic">Q</Emphasis>-learning-based swarm optimization algorithm for economic dispatch problem

In this paper, we treat optimization problems as a kind of reinforcement learning problems regarding an optimization procedure for searching an optimal solution as a reinforcement learning procedure for finding the best policy to maximize the expected rewards. This viewpoint motivated us to propose a Q-learning-based swarm optimization (QSO) algorithm. The proposed QSO algorithm is a population-based optimization algorithm which integrates the essential properties of Q-learning and particle swarm optimization. The optimization procedure of the QSO algorithm proceeds as each individual imitates the behavior of the global best one in the swarm. The best individual is chosen based on its accumulated performance instead of its momentary performance at each evaluation. Two data sets including a set of benchmark functions and a real-world problem—the economic dispatch (ED) problem for power systems—were used to test the performance of the proposed QSO algorithm. The simulation results on the benchmark functions show that the proposed QSO algorithm is comparable to or even outperforms several existing optimization algorithms. As for the ED problem, the proposed QSO algorithm has found solutions better than all previously found solutions.     

动态误分类代价下代价敏感属性选择分治算法

代价敏感属性选择问题的目的是通过权衡测试代价和误分类代价,得到一个具有最小总代价的属性子集。目前,多数代价敏感属性选择方法只考虑误分类代价固定不变的情况,不能较好地解决类分布不均衡等问题。而在大规模数据集上,算法效率不理想也是代价敏感属性选择的主要问题之一。针对这些问题,以总代价最小为目标,设计了一种新的动态误分类代价机制。结合分治思想,根据数据集规模按列自适应拆分各数据集。基于动态误分类代价重新定义最小代价属性选择问题,提出了动态误分类代价下的代价敏感属性选择分治算法。通过实验表明,该算法能在提高效率的同时获得最优误分类代价,从而保证所得属性子集的总代价最小。     

Irrevocable-choice algorithms for sampling from a stream

The problem of sampling from data streams has attracted significant interest in the last decade. Whichever sampling criteria is considered (uniform sample, maximally diverse sample, etc.), the challenges stem from the relatively small amount of memory available in the face of unbounded streams. In this work we consider an interesting extension of this problem, the framework of which is stimulated by recent improvements in sensing technologies and robotics. In some situations it is not only possible to digitally sense some aspects of the world, but to physically capture a tangible aspect of that world. Currently deployed examples include devices that can capture water/air samples, and devices that capture individual insects or fish. Such devices create an interesting twist on the stream sampling problem, because in most cases, the decision to take a physical sample is irrevocable. In this work we show how to generalize diversification sampling strategies to the irrevocable-choice setting, demonstrating our ideas on several real world domains.