Showing papers in "IEEE Transactions on Knowledge and Data Engineering in 2009"

Learning from Imbalanced Data

Abstract: With the continuous expansion of data availability in many large-scale, complex, and networked systems, such as surveillance, security, Internet, and finance, it becomes critical to advance the fundamental understanding of knowledge discovery and analysis from raw data to support decision-making processes. Although existing knowledge discovery and data engineering techniques have shown great success in many real-world applications, the problem of learning from imbalanced data (the imbalanced learning problem) is a relatively new challenge that has attracted growing attention from both academia and industry. The imbalanced learning problem is concerned with the performance of learning algorithms in the presence of underrepresented data and severe class distribution skews. Due to the inherent complex characteristics of imbalanced data sets, learning from such data requires new understandings, principles, algorithms, and tools to transform vast amounts of raw data efficiently into information and knowledge representation. In this paper, we provide a comprehensive review of the development of research in learning from imbalanced data. Our focus is to provide a critical review of the nature of the problem, the state-of-the-art technologies, and the current assessment metrics used to evaluate learning performance under the imbalanced learning scenario. Furthermore, in order to stimulate future research in this field, we also highlight the major opportunities and challenges, as well as potential important research directions for learning from imbalanced data.

Efficient Tree Structures for High Utility Pattern Mining in Incremental Databases

Abstract: Recently, high utility pattern (HUP) mining is one of the most important research issues in data mining due to its ability to consider the nonbinary frequency values of items in transactions and different profit values for every item. On the other hand, incremental and interactive data mining provide the ability to use previous data structures and mining results in order to reduce unnecessary calculations when a database is updated, or when the minimum threshold is changed. In this paper, we propose three novel tree structures to efficiently perform incremental and interactive HUP mining. The first tree structure, Incremental HUP Lexicographic Tree (IHUPL-Tree), is arranged according to an item's lexicographic order. It can capture the incremental data without any restructuring operation. The second tree structure is the IHUP transaction frequency tree (IHUPTF-Tree), which obtains a compact size by arranging items according to their transaction frequency (descending order). To reduce the mining time, the third tree, IHUP-transaction-weighted utilization tree (IHUPTWU-Tree) is designed based on the TWU value of items in descending order. Extensive performance analyses show that our tree structures are very efficient and scalable for incremental and interactive HUP mining.

A Survey of Uncertain Data Algorithms and Applications

Abstract: In recent years, a number of indirect data collection methodologies have lead to the proliferation of uncertain data. Such data points are often represented in the form of a probabilistic function, since the corresponding deterministic value is not known. This increases the challenge of mining and managing uncertain data, since the precise behavior of the underlying data is no longer known. In this paper, we provide a survey of uncertain data mining and management applications. In the field of uncertain data management, we will examine traditional methods such as join processing, query processing, selectivity estimation, OLAP queries, and indexing. In the field of uncertain data mining, we will examine traditional mining problems such as classification and clustering. We will also examine a general transform based technique for mining uncertain data. We discuss the models for uncertain data, and how they can be leveraged in a variety of applications. We discuss different methodologies to process and mine uncertain data in a variety of forms.

Unsupervised Multiway Data Analysis: A Literature Survey

Abstract: Two-way arrays or matrices are often not enough to represent all the information in the data and standard two-way analysis techniques commonly applied on matrices may fail to find the underlying structures in multi-modal datasets. Multiway data analysis has recently become popular as an exploratory analysis tool in discovering the structures in higher-order datasets, where data have more than two modes. We provide a review of significant contributions in the literature on multiway models, algorithms as well as their applications in diverse disciplines including chemometrics, neuroscience, social network analysis, text mining and computer vision.

RiMOM: A Dynamic Multistrategy Ontology Alignment Framework

Abstract: Ontology alignment identifies semantically matching entities in different ontologies. Various ontology alignment strategies have been proposed; however, few systems have explored how to automatically combine multiple strategies to improve the matching effectiveness. This paper presents a dynamic multistrategy ontology alignment framework, named RiMOM. The key insight in this framework is that similarity characteristics between ontologies may vary widely. We propose a systematic approach to quantitatively estimate the similarity characteristics for each alignment task and propose a strategy selection method to automatically combine the matching strategies based on two estimated factors. In the approach, we consider both textual and structural characteristics of ontologies. With RiMOM, we participated in the 2006 and 2007 campaigns of the Ontology Alignment Evaluation Initiative (OAEI). Our system is among the top three performers in benchmark data sets.

Patch Alignment for Dimensionality Reduction

Abstract: Spectral analysis-based dimensionality reduction algorithms are important and have been popularly applied in data mining and computer vision applications. To date many algorithms have been developed, e.g., principal component analysis, locally linear embedding, Laplacian eigenmaps, and local tangent space alignment. All of these algorithms have been designed intuitively and pragmatically, i.e., on the basis of the experience and knowledge of experts for their own purposes. Therefore, it will be more informative to provide a systematic framework for understanding the common properties and intrinsic difference in different algorithms. In this paper, we propose such a framework, named "patch alignment,rdquo which consists of two stages: part optimization and whole alignment. The framework reveals that (1) algorithms are intrinsically different in the patch optimization stage and (2) all algorithms share an almost identical whole alignment stage. As an application of this framework, we develop a new dimensionality reduction algorithm, termed discriminative locality alignment (DLA), by imposing discriminative information in the part optimization stage. DLA can (1) attack the distribution nonlinearity of measurements; (2) preserve the discriminative ability; and (3) avoid the small-sample-size problem. Thorough empirical studies demonstrate the effectiveness of DLA compared with representative dimensionality reduction algorithms.

Granular Computing and Knowledge Reduction in Formal Contexts

Abstract: Granular computing and knowledge reduction are two basic issues in knowledge representation and data mining. Granular structure of concept lattices with application in knowledge reduction in formal concept analysis is examined in this paper. Information granules and their properties in a formal context are first discussed. Concepts of a granular consistent set and a granular reduct in the formal context are then introduced. Discernibility matrices and Boolean functions are, respectively, employed to determine granular consistent sets and calculate granular reducts in formal contexts. Methods of knowledge reduction in a consistent formal decision context are also explored. Finally, knowledge hidden in such a context is unraveled in the form of compact implication rules.

A Novel Bayes Model: Hidden Naive Bayes

Abstract: Because learning an optimal Bayesian network classifier is an NP-hard problem, learning-improved naive Bayes has attracted much attention from researchers. In this paper, we summarize the existing improved algorithms and propose a novel Bayes model: hidden naive Bayes (HNB). In HNB, a hidden parent is created for each attribute which combines the influences from all other attributes. We experimentally test HNB in terms of classification accuracy, using the 36 UCI data sets selected by Weka, and compare it to naive Bayes (NB), selective Bayesian classifiers (SBC), naive Bayes tree (NBTree), tree-augmented naive Bayes (TAN), and averaged one-dependence estimators (AODE). The experimental results show that HNB significantly outperforms NB, SBC, NBTree, TAN, and AODE. In many data mining applications, an accurate class probability estimation and ranking are also desirable. We study the class probability estimation and ranking performance, measured by conditional log likelihood (CLL) and the area under the ROC curve (AUC), respectively, of naive Bayes and its improved models, such as SBC, NBTree, TAN, and AODE, and then compare HNB to them in terms of CLL and AUC. Our experiments show that HNB also significantly outperforms all of them.

Clustering and Sequential Pattern Mining of Online Collaborative Learning Data

Abstract: Group work is widespread in education. The growing use of online tools supporting group work generates huge amounts of data. We aim to exploit this data to support mirroring: presenting useful high-level views of information about the group, together with desired patterns characterizing the behavior of strong groups. The goal is to enable the groups and their facilitators to see relevant aspects of the group's operation and provide feedback if these are more likely to be associated with positive or negative outcomes and indicate where the problems are. We explore how useful mirror information can be extracted via a theory-driven approach and a range of clustering and sequential pattern mining. The context is a senior software development project where students use the collaboration tool TRAC. We extract patterns distinguishing the better from the weaker groups and get insights in the success factors. The results point to the importance of leadership and group interaction, and give promising indications if they are occurring. Patterns indicating good individual practices were also identified. We found that some key measures can be mined from early data. The results are promising for advising groups at the start and early identification of effective and poor practices, in time for remediation.

Communities and Emerging Semantics in Semantic Link Network: Discovery and Learning

Abstract: The World Wide Web provides plentiful contents for Web-based learning, but its hyperlink-based architecture connects Web resources for browsing freely rather than for effective learning. To support effective learning, an e-learning system should be able to discover and make use of the semantic communities and the emerging semantic relations in a dynamic complex network of learning resources. Previous graph-based community discovery approaches are limited in ability to discover semantic communities. This paper first suggests the semantic link network (SLN), a loosely coupled semantic data model that can semantically link resources and derive out implicit semantic links according to a set of relational reasoning rules. By studying the intrinsic relationship between semantic communities and the semantic space of SLN, approaches to discovering reasoning-constraint, rule-constraint, and classification-constraint semantic communities are proposed. Further, the approaches, principles, and strategies for discovering emerging semantics in dynamic SLNs are studied. The basic laws of the semantic link network motion are revealed for the first time. An e-learning environment incorporating the proposed approaches, principles, and strategies to support effective discovery and learning is suggested.

URBE: Web Service Retrieval Based on Similarity Evaluation

Abstract: In this work, we present UDDI registry by example (Urbe), a novel approach for Web service retrieval based on the evaluation of similarity between Web service interfaces. Our approach assumes that the Web service interfaces are defined with Web service description language (WSDL) and the algorithm combines the analysis of their structures and the analysis of the terms used inside them. The higher the similarity, the less are the differences among their interfaces. As a consequence, Urbe is useful when we need to find a Web service suitable to replace an existing one that fails. Especially in autonomic systems, this situation is very common since we need to ensure the self-management, the self-configuration, the self-optimization, the self-healing, and the self-protection of the application that is based on the failed Web service. A semantic-oriented variant of the approach is also proposed, where we take advantage of annotations semantically enriching WSDL specifications. Semantic Annotation for WSDL (SAWSDL) is adopted as a language to annotate a WSDL description. The Urbe approach has been implemented in a prototype that extends a universal description, discovery and integration (UDDI) compliant Web service registry.

A Novel Density-Based Clustering Framework by Using Level Set Method

Abstract: In this paper, a new density-based clustering framework is proposed by adopting the assumption that the cluster centers in data space can be regarded as target objects in image space. First, the level set evolution is adopted to find an approximation of cluster centers by using a new initial boundary formation scheme. Accordingly, three types of initial boundaries are defined so that each of them can evolve to approach the cluster centers in different ways. To avoid the long iteration time of level set evolution in data space, an efficient termination criterion is presented to stop the evolution process in the circumstance that no more cluster centers can be found. Then, a new effective density representation called level set density (LSD) is constructed from the evolution results. Finally, the valley seeking clustering is used to group data points into corresponding clusters based on the LSD. The experiments on some synthetic and real data sets have demonstrated the efficiency and effectiveness of the proposed clustering framework. The comparisons with DBSCAN method, OPTICS method, and valley seeking clustering method further show that the proposed framework can successfully avoid the overfitting phenomenon and solve the confusion problem of cluster boundary points and outliers.

Toward a Fuzzy Domain Ontology Extraction Method for Adaptive e-Learning

Abstract: With the widespread applications of electronic learning (e-Learning) technologies to education at all levels, increasing number of online educational resources and messages are generated from the corresponding e-Learning environments. Nevertheless, it is quite difficult, if not totally impossible, for instructors to read through and analyze the online messages to predict the progress of their students on the fly. The main contribution of this paper is the illustration of a novel concept map generation mechanism which is underpinned by a fuzzy domain ontology extraction algorithm. The proposed mechanism can automatically construct concept maps based on the messages posted to online discussion forums. By browsing the concept maps, instructors can quickly identify the progress of their students and adjust the pedagogical sequence on the fly. Our initial experimental results reveal that the accuracy and the quality of the automatically generated concept maps are promising. Our research work opens the door to the development and application of intelligent software tools to enhance e-Learning.

Decompositional Rule Extraction from Support Vector Machines by Active Learning

Abstract: Support vector machines (SVMs) are currently state-of-the-art for the classification task and, generally speaking, exhibit good predictive performance due to their ability to model nonlinearities. However, their strength is also their main weakness, as the generated nonlinear models are typically regarded as incomprehensible black-box models. In this paper, we propose a new active learning-based approach (ALBA) to extract comprehensible rules from opaque SVM models. Through rule extraction, some insight is provided into the logics of the SVM model. ALBA extracts rules from the trained SVM model by explicitly making use of key concepts of the SVM: the support vectors, and the observation that these are typically close to the decision boundary. Active learning implies the focus on apparent problem areas, which for rule induction techniques are the regions close to the SVM decision boundary where most of the noise is found. By generating extra data close to these support vectors that are provided with a class label by the trained SVM model, rule induction techniques are better able to discover suitable discrimination rules. This performance increase, both in terms of predictive accuracy as comprehensibility, is confirmed in our experiments where we apply ALBA on several publicly available data sets.

A Taxonomy of Similarity Mechanisms for Case-Based Reasoning

Abstract: Assessing the similarity between cases is a key aspect of the retrieval phase in case-based reasoning (CBR). In most CBR work, similarity is assessed based on feature value descriptions of cases using similarity metrics, which use these feature values. In fact, it might be said that this notion of a feature value representation is a defining part of the CBR worldview-it underpins the idea of a problem space with cases located relative to each other in this space. Recently, a variety of similarity mechanisms have emerged that are not founded on this feature space idea. Some of these new similarity mechanisms have emerged in CBR research and some have arisen in other areas of data analysis. In fact, research on kernel-based learning is a rich source of novel similarity representations because of the emphasis on encoding domain knowledge in the kernel function. In this paper, we present a taxonomy that organizes these new similarity mechanisms and more established similarity mechanisms in a coherent framework.

Approximate Distributed K-Means Clustering over a Peer-to-Peer Network

Abstract: Data intensive peer-to-peer (P2P) networks are finding increasing number of applications. Data mining in such P2P environments is a natural extension. However, common monolithic data mining architectures do not fit well in such environments since they typically require centralizing the distributed data which is usually not practical in a large P2P network. Distributed data mining algorithms that avoid large-scale synchronization or data centralization offer an alternate choice. This paper considers the distributed K-means clustering problem where the data and computing resources are distributed over a large P2P network. It offers two algorithms which produce an approximation of the result produced by the standard centralized K-means clustering algorithm. The first is designed to operate in a dynamic P2P network that can produce clusterings by ldquolocalrdquo synchronization only. The second algorithm uses uniformly sampled peers and provides analytical guarantees regarding the accuracy of clustering on a P2P network. Empirical results show that both the algorithms demonstrate good performance compared to their centralized counterparts at the modest communication cost.

Multirelational k-Anonymity

Abstract: k-anonymity protects privacy by ensuring that data cannot be linked to a single individual. In a k-anonymous data set, any identifying information occurs in at least k tuples. Much research has been done to modify a single-table data set to satisfy anonymity constraints. This paper extends the definitions of k-anonymity to multiple relations and shows that previously proposed methodologies either fail to protect privacy or overly reduce the utility of the data in a multiple relation setting. We also propose two new clustering algorithms to achieve multirelational anonymity. Experiments show the effectiveness of the approach in terms of utility and efficiency.

A Relation-Based Page Rank Algorithm for Semantic Web Search Engines

Abstract: With the tremendous growth of information available to end users through the Web, search engines come to play ever a more critical role. Nevertheless, because of their general-purpose approach, it is always less uncommon that obtained result sets provide a burden of useless pages. The next-generation Web architecture, represented by the Semantic Web, provides the layered architecture possibly allowing overcoming this limitation. Several search engines have been proposed, which allow increasing information retrieval accuracy by exploiting a key content of semantic Web resources, that is, relations. However, in order to rank results, most of the existing solutions need to work on the whole annotated knowledge base. In this paper, we propose a relation-based page rank algorithm to be used in conjunction with semantic Web search engines that simply relies on information that could be extracted from user queries and on annotated resources. Relevance is measured as the probability that a retrieved resource actually contains those relations whose existence was assumed by the user at the time of query definition.

Nonlinear Dimensionality Reduction with Local Spline Embedding

Abstract: This paper presents a new algorithm for nonlinear dimensionality reduction (NLDR). Our algorithm is developed under the conceptual framework of compatible mapping. Each such mapping is a compound of a tangent space projection and a group of splines. Tangent space projection is estimated at each data point on the manifold, through which the data point itself and its neighbors are represented in tangent space with local coordinates. Splines are then constructed to guarantee that each of the local coordinates can be mapped to its own single global coordinate with respect to the underlying manifold. Thus, the compatibility between local alignments is ensured. In such a work setting, we develop an optimization framework based on reconstruction error analysis, which can yield a global optimum. The proposed algorithm is also extended to embed out of samples via spline interpolation. Experiments on toy data sets and real-world data sets illustrate the validity of our method.

Exact Knowledge Hiding through Database Extension

Abstract: In this paper, we propose a novel, exact border-based approach that provides an optimal solution for the hiding of sensitive frequent itemsets by (i) minimally extending the original database by a synthetically generated database part - the database extension, (ii) formulating the creation of the database extension as a constraint satisfaction problem, (iii) mapping the constraint satisfaction problem to an equivalent binary integer programming problem, (iv) exploiting underutilized synthetic transactions to proportionally increase the support of non-sensitive itemsets, (v) minimally relaxing the constraint satisfaction problem to provide an approximate solution close to the optimal one when an ideal solution does not exist, and (vi) by using a partitioning in the universe of the items to increase the efficiency of the proposed hiding algorithm. Extending the original database for sensitive itemset hiding is proved to provide optimal solutions to an extended set of hiding problems compared to previous approaches and to provide solutions of higher quality. Moreover, the application of binary integer programming enables the simultaneous hiding of the sensitive itemsets and thus allows for the identification of globally optimal solutions.

Rough Cluster Quality Index Based on Decision Theory

Abstract: Quality of clustering is an important issue in application of clustering techniques. Most traditional cluster validity indices are geometry-based cluster quality measures. This paper proposes a cluster validity index based on the decision-theoretic rough set model by considering various loss functions. Experiments with synthetic, standard, and real-world retail data show the usefulness of the proposed validity index for the evaluation of rough and crisp clustering. The measure is shown to help determine optimal number of clusters, as well as an important parameter called threshold in rough clustering. The experiments with a promotional campaign for the retail data illustrate the ability of the proposed measure to incorporate financial considerations in evaluating quality of a clustering scheme. This ability to deal with monetary values distinguishes the proposed decision-theoretic measure from other distance-based measures. The proposed validity index can also be extended for evaluating other clustering algorithms such as fuzzy clustering.

A Divide-and-Conquer Approach for Minimum Spanning Tree-Based Clustering

Abstract: Due to their ability to detect clusters with irregular boundaries, minimum spanning tree-based clustering algorithms have been widely used in practice. However, in such clustering algorithms, the search for nearest neighbor in the construction of minimum spanning trees is the main source of computation and the standard solutions take O(N2) time. In this paper, we present a fast minimum spanning tree-inspired clustering algorithm, which, by using an efficient implementation of the cut and the cycle property of the minimum spanning trees, can have much better performance than O(N2).

Distributional Features for Text Categorization

Abstract: Text categorization is the task of assigning predefined categories to natural language text. With the widely used 'bag of words' representation, previous researches usually assign a word with values such that whether this word appears in the document concerned or how frequently this word appears. Although these values are useful for text categorization, they have not fully expressed the abundant information contained in the document. This paper explores the effect of other types of values, which express the distribution of a word in the document. These novel values assigned to a word are called distributional features, which include the compactness of the appearances of the word and the position of the first appearance of the word. The proposed distributional features are exploited by a tf idf style equation and different features are combined using ensemble learning techniques. Experiments show that the distributional features are useful for text categorization. In contrast to using the traditional term frequency values solely, including the distributional features requires only a little additional cost, while the categorization performance can be significantly improved. Further analysis shows that the distributional features are especially useful when documents are long and the writing style is casual.

Predictive Ensemble Pruning by Expectation Propagation

Abstract: An ensemble is a group of learners that work together as a committee to solve a problem. The existing ensemble learning algorithms often generate unnecessarily large ensembles, which consume extra computational resource and may degrade the generalization performance. Ensemble pruning algorithms aim to find a good subset of ensemble members to constitute a small ensemble, which saves the computational resource and performs as well as, or better than, the unpruned ensemble. This paper introduces a probabilistic ensemble pruning algorithm by choosing a set of ldquosparserdquo combination weights, most of which are zeros, to prune the ensemble. In order to obtain the set of sparse combination weights and satisfy the nonnegative constraint of the combination weights, a left-truncated, nonnegative, Gaussian prior is adopted over every combination weight. Expectation propagation (EP) algorithm is employed to approximate the posterior estimation of the weight vector. The leave-one-out (LOO) error can be obtained as a by-product in the training of EP without extra computation and is a good indication for the generalization error. Therefore, the LOO error is used together with the Bayesian evidence for model selection in this algorithm. An empirical study on several regression and classification benchmark data sets shows that our algorithm utilizes far less component learners but performs as well as, or better than, the unpruned ensemble. Our results are very competitive compared with other ensemble pruning algorithms.

k-Anonymization with Minimal Loss of Information

Abstract: The technique of k-anonymization allows the releasing of databases that contain personal information while ensuring some degree of individual privacy. Anonymization is usually performed by generalizing database entries. We formally study the concept of generalization, and propose three information-theoretic measures for capturing the amount of information that is lost during the anonymization process. The proposed measures are more general and more accurate than those that were proposed by Meyerson and Williams and Aggarwal et al. We study the problem of achieving k-anonymity with minimal loss of information. We prove that it is NP-hard and study polynomial approximations for the optimal solution. Our first algorithm gives an approximation guarantee of O(ln k) for two of our measures as well as for the previously studied measures. This improves the best-known O(k)-approximation in. While the previous approximation algorithms relied on the graph representation framework, our algorithm relies on a novel hypergraph representation that enables the improvement in the approximation ratio from O(k) to O(ln k). As the running time of the algorithm is O(n2k}), we also show how to adapt the algorithm in in order to obtain an O(k)-approximation algorithm that is polynomial in both n and k.

Adapted One-versus-All Decision Trees for Data Stream Classification

Abstract: One versus all (OVA) decision trees learn k individual binary classifiers, each one to distinguish the instances of a single class from the instances of all other classes. Thus OVA is different from existing data stream classification schemes whose majority use multiclass classifiers, each one to discriminate among all the classes. This paper advocates some outstanding advantages of OVA for data stream classification. First, there is low error correlation and hence high diversity among OVA's component classifiers, which leads to high classification accuracy. Second, OVA is adept at accommodating new class labels that often appear in data streams. However, there also remain many challenges to deploy traditional OVA for classifying data streams. First, as every instance is fed to all component classifiers, OVA is known as an inefficient model. Second, OVA's classification accuracy is adversely affected by the imbalanced class distribution in data streams. This paper addresses those key challenges and consequently proposes a new OVA scheme that is adapted for data stream classification. Theoretical analysis and empirical evidence reveal that the adapted OVA can offer faster training, faster updating and higher classification accuracy than many existing popular data stream classification algorithms.

Automatically Determining the Number of Clusters in Unlabeled Data Sets

Abstract: Clustering is a popular tool for exploratory data analysis. One of the major problems in cluster analysis is the determination of the number of clusters in unlabeled data, which is a basic input for most clustering algorithms. In this paper we investigate a new method called DBE (dark block extraction) for automatically estimating the number of clusters in unlabeled data sets, which is based on an existing algorithm for visual assessment of cluster tendency (VAT) of a data set, using several common image and signal processing techniques. Basic steps include: 1) generating a VAT image of an input dissimilarity matrix; 2) performing image segmentation on the VAT image to obtain a binary image, followed by directional morphological filtering; 3) applying a distance transform to the filtered binary image and projecting the pixel values onto the main diagonal axis of the image to form a projection signal; 4) smoothing the projection signal, computing its first-order derivative, and then detecting major peaks and valleys in the resulting signal to decide the number of clusters. Our new DBE method is nearly "automatic", depending on just one easy-to-set parameter. Several numerical and real-world examples are presented to illustrate the effectiveness of DBE.

Development of a Software Engineering Ontology for Multisite Software Development

Abstract: This paper aims to present an ontology model of software engineering to represent its knowledge. The fundamental knowledge relating to software engineering is well described in the textbook entitled Software Engineering by Sommerville that is now in its eighth edition (2004) and the white paper, Software Engineering Body of Knowledge (SWEBOK), by the IEEE (203) upon which software engineering ontology is based. This paper gives an analysis of what software engineering ontology is, what it consists of, and what it is used for in the form of usage example scenarios. The usage scenarios presented in this paper highlight the characteristics of the software engineering ontology. The software engineering ontology assists in defining information for the exchange of semantic project information and is used as a communication framework. Its users are software engineers sharing domain knowledge as well as instance knowledge of software engineering.

Nonlinear Dimension Reduction with Kernel Sliced Inverse Regression

Abstract: Sliced inverse regression (SIR) is a renowned dimension reduction method for finding an effective low-dimensional linear subspace. Like many other linear methods, SIR can be extended to nonlinear setting via the ldquokernel trick.rdquo The main purpose of this paper is two-fold. We build kernel SIR in a reproducing kernel Hilbert space rigorously for a more intuitive model explanation and theoretical development. The second focus is on the implementation algorithm of kernel SIR for fast computation and numerical stability. We adopt a low-rank approximation to approximate the huge and dense full kernel covariance matrix and a reduced singular value decomposition technique for extracting kernel SIR directions. We also explore kernel SIR's ability to combine with other linear learning algorithms for classification and regression including multiresponse regression. Numerical experiments show that kernel SIR is an effective kernel tool for nonlinear dimension reduction and it can easily combine with other linear algorithms to form a powerful toolkit for nonlinear data analysis.

Batch Mode Active Learning with Applications to Text Categorization and Image Retrieval

Abstract: Most machine learning tasks in data classification and information retrieval require manually labeled data examples in the training stage. The goal of active learning is to select the most informative examples for manual labeling in these learning tasks. Most of the previous studies in active learning have focused on selecting a single unlabeled example in each iteration. This could be inefficient, since the classification model has to be retrained for every acquired labeled example. It is also inappropriate for the setup of information retrieval tasks where the user's relevance feedback is often provided for the top K retrieved items. In this paper, we present a framework for batch mode active learning, which selects a number of informative examples for manual labeling in each iteration. The key feature of batch mode active learning is to reduce the redundancy among the selected examples such that each example provides unique information for model updating. To this end, we employ the Fisher information matrix as the measurement of model uncertainty, and choose the set of unlabeled examples that can efficiently reduce the Fisher information of the classification model. We apply our batch mode active learning framework to both text categorization and image retrieval. Promising results show that our algorithms are significantly more effective than the active learning approaches that select unlabeled examples based only on their informativeness for the classification model.