Partial example acquisition in cost-sensitive learning

Imbalanced data classification is one of the challenging problems in data mining and machine learning research. The traditional classification algorithms are often biased towards the majority class when learning from imbalanced data. Much work have been proposed to address this problem, including data re-sampling, algorithm modification, and cost-sensitive learning. However, most of them focus on one of these techniques. This paper proposes to utilize both algorithm modification and cost-sensitive learning based on decision-theoretic rough set (DTRS) model. In particular, we use naive Bayes classifier as the base classifier and modify it for imbalanced learning. For cost-sensitive learning, we adopt the systematic method from DTRS to derive required thresholds that have the minimum decision cost. Our experimental results on three well-known text classification databases show that unified DTRS provides similar performance on balanced class distribution, outperforms naive Bayes classifier on imbalanced datasets, and is competitive with other imbalanced learning classifier.

In domains in which single agent learning is a more natural metaphor for an artifact-embedded agent, Exemplar-Based Learning (EBL) requires significantly large sets of training examples for it to be applicable. Obviously large sets of training examples contradict resource capabilities of artifacts. To make EBL a possibility for these artifacts, sets of training examples must be reduced in size in a way that does not compromise learning performance in order to relieve artifacts' resources (e.g. memory). In this paper, we investigate training sets requirements for artifacts learning and propose a ranking-based Stratified Ordered Selection (SOS) method to scale them down. Contrary to reduction approaches in mainstream learning, this method has been designed with resource constraint nature of artifacts in mind. Artifacts shall use an intermediary which implements SOS to, dynamically and on-demand, retrieve training subsets based on their resource capacities (e.g. memory, CPU). SOS uses a new Level Order (LO) ranking scheme which has been designed to broaden representation of classes of examples, to quicken data retrieval, and to allow for retrieval of subsets of varying sizes while ensuring same or near same learning performance. We present how SOS evaluates on various well known machine learning datasets and how it compares to some of the best performing data reduction approaches.

Research in machine learning, data mining and related areas has produced a wide variety of algorithms for cost-sensitive (CS) classification, where instead of maximizing the classification accuracy, minimizing the misclassification cost becomes the objective. However, these methods assume that training sets do not contain significant noise, which is rarely the case in real-world environments. In this paper, we systematically study the impacts of class noise on CS learning, and propose a cost-guided class noise handling algorithm to identify noise for effective CS learning. We call it cost-guided iterative classification filter (CICF), because it seamlessly integrates costs and an existing classification filter (C. Brodley and M. Friedl, 1999) for noise identification. Instead of putting equal weights to handle noise in all classes in existing efforts, CICF puts more emphasis on expensive classes, which makes it especially successful in dealing with datasets with a large cost-ratio. Experimental results and comparative studies from real-world datasets indicate that the existence of noise may seriously corrupt the performance of CS classifiers, and by adopting the proposed CICF algorithm, we can significantly reduce the misclassification cost of a CS classifier in noisy environments.

Data mining requires certain information---for example, supervised learning requires training data. Some prior research has recognized that this information often does not simply present itself for free, but involves various acquisition costs. In addition, applying the learned models involves costs and benefits. I introduce a general economic setting that includes as special cases the settings of many different streams of prior research, such as cost-sensitive learning, traditional active learning, semi-supervised learning, active feature acquisition, progressive sampling, and budgeted learning, which are interwoven inextricably. For data mining in the general setting I suggest a strategy of maximum expected-utility data acquisition. Finally, I discuss how there are many open research issues that must be addressed. As a simple example, we must be able to deal with the seemingly straightforward problem of handling missing values in induction and inference.

CHAPTER 14 - Representation Generation in an Exploratory Learning System

This paper introduces a method that improves the performance of a problem solver by reformulating its domain theory into one in which functionally relevant features are explicit in the syntax. This method, in contrast to previous reformulation methods, employs sets of training examples to constrain and direct the reformulation process. The use of examples offers two advantages over purely deductive approaches: First, the examples identify the exact part of the domain theory to be reformulated. Second, a proof with examples is much simpler to construct than a general proof because it is fully instantiated. The method exploits the fact that what is relevant to a goal is syntactically explicit in successful solutions to that goal. The method first takes as input a set of training examples that “exercise” an important part of the domain theory and then applies the problem solver to explain the examples in terms of a relevant goal. Next, the set of explanations is “clustered” into cases and then generalized using the induction over explanations method, forming a set of general explanations. Finally, these general explanations are reformulated into new domain theory rules. We illustrate the method in the domain of chess. We reformulate a simple declarative encoding of legal-move to produce a new domain theory that can generate the legal moves in a tenth of the time required by the original theory. We also show how the reformulated theory can more efficiently describe the important knight-fork feature.

Whoso neglects learning in his youth, loses the past and is dead for the future . –Euripides (484 BC – 406 BC), Phrixus, Frag. 927 Learning is the ability of an agent to improve its behavior based on experience. This could mean the following: The range of behaviors is expanded; the agent can do more. The accuracy on tasks is improved; the agent can do things better. The speed is improved; the agent can do things faster. The ability to learn is essential to any intelligent agent. As Euripides pointed, learning involves an agent remembering its past in a way that is useful for its future. This chapter considers supervised learning: given a set of training examples made up of input–output pairs, predict the output of a new input. We show how such learning may be based on one of four possible approaches: choosing a single hypothesis that fits the training examples well, predicting directly from the training examples, selecting the subset of a hypothesis space consistent with the training examples, or finding the posterior probability distribution of hypotheses conditioned on the training examples. Chapter 11 goes beyond supervised learning and considers clustering (often called unsupervised learning), learning probabilistic models, and reinforcement learning. Section 14.2 (page 606) considers learning relational representations.

Who so neglects learning in his youth, loses the past and is dead for the future. – Euripides (484 BC – 406 BC), Phrixus, Frag. 927 Learning is the ability of an agent to improve its behavior based on experience. This could mean the following • The range of behaviors is expanded; the agent can do more. • The accuracy on tasks is improved; the agent can do things better. • The speed is improved; the agent can do things faster. The ability to learn is essential to any intelligent agent. As Euripides pointed out, learning involves an agent remembering its past in a way that is useful for its future. This chapter considers the problem of making a prediction as supervised learning: given a set of training examples made up of input–output pairs, predict the output of a new example where only the inputs are given. We explore four approaches to learning: choosing a single hypothesis that fits the training examples well, predicting directly from the training examples, selecting the subset of a hypothesis space consistent with the training examples, or (in Section 10.4 (page 512)) predicting based on the posterior probability distribution of hypotheses conditioned on the training examples. Chapter 10 considers learning probabilistic models. Chapter 12 covers reinforcement learning. Section 15.2 (page 701) considers learning relational representations. Learning Issues The following components are part of any learning problem: Task The behavior or task that is being improved Data The experiences that are used to improve performance in the task, usually in the form of a sequence of examples Measure of improvement How the improvement is measured – for example, new skills that were not present initially, increasing accuracy in prediction, or improved speed Consider the agent internals of Figure 2.9 (page 66). The problem of learning is to take in prior knowledge and data (e.g., about the experiences of the agent) and to create an internal representation (the knowledge base) that is used by the agent as it acts. Learning techniques face the following issues: Task Virtually any task for which an agent can get data or experiences can be learned.

Accelerating neural network training using weight extrapolations

We develop a novel framework that aims at automatically adapting previously learned information extraction knowledge from a source Web site to a new unseen target site in the same domain. Two kinds of features related to the text fragments from the Web documents are investigated. The first type of feature is called, a site-invariant feature. These features likely remain unchanged in Web pages from different sites in the same domain. The second type of feature is called a site-dependent feature. These features are different in the Web pages collected from different Web sites, while they are similar in the Web pages originating from the same site. In our framework, we derive the site-invariant features from previously learned extraction knowledge and the items previously collected or extracted from the source Web site. The derived site-invariant features will be exploited to automatically seek a new set of training examples in the new unseen target site. Both the site-dependent features and the site-invariant features of these automatically discovered training examples will be considered in the learning of new information extraction knowledge for the target site. We conducted extensive experiments on a set of real-world Web sites collected from three different domains to demonstrate the performance of our framework. For example, by just providing training examples from one online book catalog Web site, our approach can automatically extract information from ten different book catalog sites achieving an average precision and recall of 71.9% and 84.0% respectively without any further manual intervention.

A mixed-integer programming approach to GRNN parameter estimation

Sentiment analysis has been rapidly employed for business decision support. New data mining researchers are yet to have an adequate understanding of the various applications of sentiment analysis while utilising social media data. As a result, it is critical to define the data mining and text analytics research trend holistically using existing literature. The study explores sentiment analysis research for its application in transforming social media data and identifies relevant research aspects through a comprehensive bibliometric review of 523 research articles published in the Scopus database (between 2018 and 2022) to discern the content and thematic analysis. Findings suggested that key purposes of the sentiment analysis are mainly related to innovation, transparency, and efficiency. Our review also highlights the distinctiveness of sentiment analysis for synthesising social media information to investigate various features, including the knowledge-domain map that detects author collaboration networks in the past.

Learning effective and efficient classifiers for imbalanced data is one of ten challenge problems in data mining research. Studying classifiers for imbalanced data is a popular area in machine learning and data mining, which also has great significance in many areas, such as cancer diagnose, credit card fraud detection and intrusion detection. The study for imbalanced data classification can be divided into three major parts: resampling data, change internal algorithm and cost-sensitive learning. Weighted feature methods can also enhance the accuracy of classification. In this paper, we intend to use a novel method combining weighted feature and cost-sensitive learning to deal with imbalanced data. We add weights to features, which causes that the position of each instance in space changes. Our goal is to increase the separation between classes by enlarging the space around the separating boundary surface through weighted features. Since the margin is enlarged, the chance that the instances in minority class are classified into majority class by mistake will be lower. Weighted Features cost-sensitive SVM (WF-CSSVM) performs well in both accuracy and cost. UCI datasets are utilized in experiment part and most of them can be classified perfectly. Accuracy, G-mean and ROC are employed as evaluation metrics.

Data mining (DM) research has successfully developed advanced DM techniques and algorithms over the last few decades, and many organisations have great expectations to take more benefit of their data warehouses in decision making. Currently, the strong focus of most DM-researchers is still only on technology-oriented topics. Commonly the DM research has several stakeholders, the major of which can be divided into internal and external ones each having their own point of view, and which are at least partly conflicting. The most important internal groups of stakeholders are the DM research community and academics in other disciplines. The most important external stakeholder groups are managers and domain experts who have their own utility-based interests to DM and DM research results. In this paper we discuss these practice-oriented points of view towards DM research and suggest broader discussions inside the DM research community about who should do that kind of research. We bring in the discussion several topics developed in the information systems (IS) discipline and show some similarities between IS and DM systems. DM systems have also their own peculiarities and we conclude that researchers who take into account human and organisational aspects related to DM systems need to have also some understanding about DM. This makes us suggest that the research area inside the DM community should be made broader than the current heavily technology-oriented one.

Cost-sensitive matrixized classification learning with information entropy