Showing papers by "Hong Kong University of Science and Technology published in 2007"

Top 10 algorithms in data mining

Abstract: This paper presents the top 10 data mining algorithms identified by the IEEE International Conference on Data Mining (ICDM) in December 2006: C4.5, k-Means, SVM, Apriori, EM, PageRank, AdaBoost, kNN, Naive Bayes, and CART. These top 10 algorithms are among the most influential data mining algorithms in the research community. With each algorithm, we provide a description of the algorithm, discuss the impact of the algorithm, and review current and further research on the algorithm. These 10 algorithms cover classification, clustering, statistical learning, association analysis, and link mining, which are all among the most important topics in data mining research and development.

A second generation human haplotype map of over 3.1 million SNPs

Abstract: We describe the Phase II HapMap, which characterizes over 3.1 million human single nucleotide polymorphisms (SNPs) genotyped in 270 individuals from four geographically diverse populations and includes 25-35% of common SNP variation in the populations surveyed. The map is estimated to capture untyped common variation with an average maximum r2 of between 0.9 and 0.96 depending on population. We demonstrate that the current generation of commercial genome-wide genotyping products captures common Phase II SNPs with an average maximum r2 of up to 0.8 in African and up to 0.95 in non-African populations, and that potential gains in power in association studies can be obtained through imputation. These data also reveal novel aspects of the structure of linkage disequilibrium. We show that 10-30% of pairs of individuals within a population share at least one region of extended genetic identity arising from recent ancestry and that up to 1% of all common variants are untaggable, primarily because they lie within recombination hotspots. We show that recombination rates vary systematically around genes and between genes of different function. Finally, we demonstrate increased differentiation at non-synonymous, compared to synonymous, SNPs, resulting from systematic differences in the strength or efficacy of natural selection between populations.

Genome-wide detection and characterization of positive selection in human populations

Abstract: With the advent of dense maps of human genetic variation, it is now possible to detect positive natural selection across the human genome. Here we report an analysis of over 3 million polymorphisms from the International HapMap Project Phase 2 (HapMap2). We used 'long-range haplotype' methods, which were developed to identify alleles segregating in a population that have undergone recent selection, and we also developed new methods that are based on cross-population comparisons to discover alleles that have swept to near-fixation within a population. The analysis reveals more than 300 strong candidate regions. Focusing on the strongest 22 regions, we develop a heuristic for scrutinizing these regions to identify candidate targets of selection. In a complementary analysis, we identify 26 non-synonymous, coding, single nucleotide polymorphisms showing regional evidence of positive selection. Examination of these candidates highlights three cases in which two genes in a common biological process have apparently undergone positive selection in the same population:LARGE and DMD, both related to infection by the Lassa virus, in West Africa;SLC24A5 and SLC45A2, both involved in skin pigmentation, in Europe; and EDAR and EDA2R, both involved in development of hair follicles, in Asia.

Boosting for transfer learning

Abstract: Traditional machine learning makes a basic assumption: the training and test data should be under the same distribution. However, in many cases, this identical-distribution assumption does not hold. The assumption might be violated when a task from one new domain comes, while there are only labeled data from a similar old domain. Labeling the new data can be costly and it would also be a waste to throw away all the old data. In this paper, we present a novel transfer learning framework called TrAdaBoost, which extends boosting-based learning algorithms (Freund & Schapire, 1997). TrAdaBoost allows users to utilize a small amount of newly labeled data to leverage the old data to construct a high-quality classification model for the new data. We show that this method can allow us to learn an accurate model using only a tiny amount of new data and a large amount of old data, even when the new data are not sufficient to train a model alone. We show that TrAdaBoost allows knowledge to be effectively transferred from the old data to the new. The effectiveness of our algorithm is analyzed theoretically and empirically to show that our iterative algorithm can converge well to an accurate model.

Impetus for Action: A Cultural Analysis of Justice and Organizational Citizenship Behavior in Chinese Society

Abstract: To understand variations in citizenship behavior within a culture, we examine the relationship between citizenship behaviors and organizational justice in two studies in a Chinese context, using two cultural characteristics (traditionality and modernity) and one individual (gender) characteristic. In Study 1, we develop an indigenous measure of organizational citizenship behavior and explore the similarities and differences of this measure with its Western counterpart. In Study 2, we use this citizenship behavior measure to test its relationship to justice. Results demonstrate that organizational justice (distributive and procedural) is most strongly related to citizenship behavior for individuals who endorse less traditional, or high modernity, values. In addition, we found the relationship between justice and citizenship behavior to be stronger for men than for women. The studies are discussed in terms of the generality of citizenship behavior and its relation to organizational justiqe and cultural characteristics.

Correlations between Percolation Threshold, Dispersion State, and Aspect Ratio of Carbon Nanotubes

Abstract: Critical factors that determine the percolation threshold of carbon nanotube (CNT)-reinforced polymer nanocomposites are studied. An improved analytical model is developed based on an interparticle distance concept. Two dispersion parameters are introduced in the model to correctly reflect the different dispersion states of CNTs in the matrix—entangled bundles and well-dispersed individual CNTs. CNT–epoxy nanocomposites with different dispersion states are fabricated from the same constituent materials by employing different processing conditions. The corresponding percolation thresholds of the nanocomposites vary over a wide range, from 0.1 to greater than 1.0 wt %, and these variations are explained in terms of dispersion parameters and aspect ratios of CNTs. Important factors that control the percolation threshold of nanocomposites are identified based on the comparison between modeling data and experimental results.

Image deblurring with blurred/noisy image pairs

Abstract: Taking satisfactory photos under dim lighting conditions using a hand-held camera is challenging. If the camera is set to a long exposure time, the image is blurred due to camera shake. On the other hand, the image is dark and noisy if it is taken with a short exposure time but with a high camera gain. By combining information extracted from both blurred and noisy images, however, we show in this paper how to produce a high quality image that cannot be obtained by simply denoising the noisy image, or deblurring the blurred image alone. Our approach is image deblurring with the help of the noisy image. First, both images are used to estimate an accurate blur kernel, which otherwise is difficult to obtain from a single blurred image. Second, and again using both images, a residual deconvolution is proposed to significantly reduce ringing artifacts inherent to image deconvolution. Third, the remaining ringing artifacts in smooth image regions are further suppressed by a gain-controlled deconvolution process. We demonstrate the effectiveness of our approach using a number of indoor and outdoor images taken by off-the-shelf hand-held cameras in poor lighting environments.

Power Control By Geometric Programming

Abstract: In wireless cellular or ad hoc networks where Quality of Service (QoS) is interference-limited, a variety of power control problems can be formulated as nonlinear optimization with a system-wide objective, e.g., maximizing the total system throughput or the worst user throughput, subject to QoS constraints from individual users, e.g., on data rate, delay, and outage probability. We show that in the high Signal-to- interference Ratios (SIR) regime, these nonlinear and apparently difficult, nonconvex optimization problems can be transformed into convex optimization problems in the form of geometric programming; hence they can be very efficiently solved for global optimality even with a large number of users. In the medium to low SIR regime, some of these constrained nonlinear optimization of power control cannot be turned into tractable convex formulations, but a heuristic can be used to compute in most cases the optimal solution by solving a series of geometric programs through the approach of successive convex approximation. While efficient and robust algorithms have been extensively studied for centralized solutions of geometric programs, distributed algorithms have not been explored before. We present a systematic method of distributed algorithms for power control that is geometric-programming-based. These techniques for power control, together with their implications to admission control and pricing in wireless networks, are illustrated through several numerical examples.

Corporate Disclosures by Family Firms

Abstract: Compared to non-family firms, family firms face less severe agency problems due to the separation of ownership and management, but more severe agency problems that arise between controlling and non-controlling shareholders. These characteristics of family firms affect their corporate disclosure practices. For S&P 500 firms, we show that family firms report better quality earnings, are more likely to warn for a given magnitude of bad news, but make fewer disclosures about their corporate governance practices. Consistent with family firms making better financial disclosures, we find that family firms have larger analyst following, more informative analysts' forecasts, and smaller bid-ask spreads.

Role of ventilation in airborne transmission of infectious agents in the built environment - a multidisciplinary systematic review.

Abstract: There have been few recent studies demonstrating a definitive association between the transmission of airborne infections and the ventilation of buildings. The severe acute respiratory syndrome (SARS) epidemic in 2003 and current concerns about the risk of an avian influenza (H5N1) pandemic, have made a review of this area timely. We searched the major literature databases between 1960 and 2005, and then screened titles and abstracts, and finally selected 40 original studies based on a set of criteria. We established a review panel comprising medical and engineering experts in the fields of microbiology, medicine, epidemiology, indoor air quality, building ventilation, etc. Most panel members had experience with research into the 2003 SARS epidemic. The panel systematically assessed 40 original studies through both individual assessment and a 2-day face-to-face consensus meeting. Ten of 40 studies reviewed were considered to be conclusive with regard to the association between building ventilation and the transmission of airborne infection. There is strong and sufficient evidence to demonstrate the association between ventilation, air movements in buildings and the transmission/spread of infectious diseases such as measles, tuberculosis, chickenpox, influenza, smallpox and SARS. There is insufficient data to specify and quantify the minimum ventilation requirements in hospitals, schools, offices, homes and isolation rooms in relation to spread of infectious diseases via the airborne route. PRACTICAL IMPLICATION: The strong and sufficient evidence of the association between ventilation, the control of airflow direction in buildings, and the transmission and spread of infectious diseases supports the use of negatively pressurized isolation rooms for patients with these diseases in hospitals, in addition to the use of other engineering control methods. However, the lack of sufficient data on the specification and quantification of the minimum ventilation requirements in hospitals, schools and offices in relation to the spread of airborne infectious diseases, suggest the existence of a knowledge gap. Our study reveals a strong need for a multidisciplinary study in investigating disease outbreaks, and the impact of indoor air environments on the spread of airborne infectious diseases.

Individual-Level Cultural Values as Moderators of Perceived Organizational Support–Employee Outcome Relationships in China: Comparing the Effects of Power Distance and Traditionality

Abstract: Drawing on a cross-organizational sample of 163 supervisor-subordinate dyads from mainland China, we examined the moderating effect of power distance and Chinese traditionality on relationships bet...

Trade-Off and Pecking Order Theories of Debt

Abstract: Taxes, bankruptcy costs, transactions costs, adverse selection, and agency conflicts have all been advocated as major explanations for the corporate use of debt financing. These ideas have often been synthesized into the trade-off theory and the pecking order theory of leverage. These theories and the related evidence are reviewed in this survey. A number of important empirical stylized facts are identified. To understand the evidence, it is important to recognize the differences among private firms, small public firms and large public firms. Private firms seem to use retained earnings and bank debt heavily. Small public firms make active use of equity financing. Large public firms primarily use retained earnings and corporate bonds. The available evidence can be interpreted in several ways. Direct transaction costs and indirect bankruptcy costs appear to play important roles in a firm's choice of debt. The relative importance of the other factors remains open to debate. No currently available model appears capable of simultaneously accounting for all of the stylized facts.

Manipulating Electromagnetic Wave Polarizations by Anisotropic Metamaterials

Abstract: We show that the polarization states of electromagnetic waves can be manipulated through reflections by an anisotropic metamaterial plate, and all possible polarizations (circular, elliptic, and linear) are realizable via adjusting material parameters. In particular, a linearly polarized light converts its polarization completely to the cross direction after reflection under certain conditions. Microwave experiments were performed to successfully realize these ideas and results are in excellent agreement with numerical simulations.

An empirical examination of the mechanisms mediating between high-performance work systems and the performance of Japanese organizations.

Abstract: The resource-based view of the firm and social exchange perspectives are invoked to hypothesize linkages among high-performance work systems, collective human capital, the degree of social exchange in an establishment, and establishment performance. The authors argue that high-performance work systems generate a high level of collective human capital and encourage a high degree of social exchange within an organization, and that these are positively related to the organization's overall performance. On the basis of a sample of Japanese establishments, the results provide support for the existence of these mediating mechanisms through which high-performance work systems affect overall establishment performance.

Preventing Location-Based Identity Inference in Anonymous Spatial Queries

Abstract: The increasing trend of embedding positioning capabilities (for example, GPS) in mobile devices facilitates the widespread use of location-based services. For such applications to succeed, privacy and confidentiality are essential. Existing privacy-enhancing techniques rely on encryption to safeguard communication channels, and on pseudonyms to protect user identities. Nevertheless, the query contents may disclose the physical location of the user. In this paper, we present a framework for preventing location-based identity inference of users who issue spatial queries to location-based services. We propose transformations based on the well-established K-anonymity concept to compute exact answers for range and nearest neighbor search, without revealing the query source. Our methods optimize the entire process of anonymizing the requests and processing the transformed spatial queries. Extensive experimental studies suggest that the proposed techniques are applicable to real-life scenarios with numerous mobile users.

The Influence of Relational Demography and Guanxi: The Chinese Case

Abstract: Two studies were conducted to investigate the indigenous concept of guanxi and its applications in the Chinese context. Guanxi refers to the existence of direct particularistic ties between an individual and others. We relate the concept to the idea of relational demography, which refers to similarities or differences between an individual and others on such factors as age, gender, race, religion, education, and occupation. The two studies focused on the importance of guanxi and relational demography in Chinese employment settings. In study 1, their importance was examined in a sample of 560 vertical dyads (i.e., between supervisor and subordinate) in Taiwan. In study 2, the effects were analyzed in a sample of 205 horizontal dyads, specifically between business executives and their important business connections (e.g., key customers, suppliers, bankers, government officials) in mainland China. Results support the importance of both guanxi and relational demography for subordinate trust in the supervisor, but only guanxi is found to be (extremely) important for business executives' trust in their connections. Implications for future cross-cultural research on the effect of common ties are discussed.

Defect emissions in ZnO nanostructures

Abstract: Defects in three different types of ZnO nanostructures before and after annealing under different conditions were studied. The annealing atmosphere and temperature were found to strongly affect the yellow and orange-red defect emissions, while green emission was not significantly affected by annealing. The defect emissions exhibited a strong dependence on the temperature and excitation wavelength, with some defect emissions observable only at low temperatures and for certain excitation wavelengths. The yellow emission in samples prepared by a hydrothermal method is likely due to the presence of OH groups, instead of the commonly assumed interstitial oxygen defect. The green and orange-red emissions are likely due to donor acceptor transitions involving defect complexes, which likely include zinc vacancy complexes in the case of orange-red emissions.

Reduction of Mutual Coupling Between Closely-Packed Antenna Elements

Abstract: A simple ground plane structure that can reduce mutual coupling between closely-packed antenna elements is proposed and studied. The structure consists of a slitted pattern, without via's, etched onto a single ground plane and it is therefore low cost and straightforward to fabricate. It is found that isolations of more than -20 dB can be achieved between two parallel individual planar inverted-F antennas (PIFAs) sharing a common ground plane, with inter-antenna spacing (center to center) of 0.116 lambdao and ground plane size 0.331lambdao 2. At 2.31 GHz it is demonstrated that this translates into an edge to edge separation between antennas of just 10 mm. Similarly the structure can be applied to reduce mutual coupling between three or four radiating elements. In addition the mutual coupling between half wavelength patches and monopoles can also be reduced with the aid of the proposed ground plane structure. Results of parametric studies are also given in this paper. Both simulation and measurement results are used to confirm the suppression of mutual coupling between closely-packed antenna elements with our slitted ground plane.

Foreign Entry and Survival: Effects of Strategic Choices on Performance in International Markets

Abstract: This paper investigates effective strategies that can reduce the risk of failure in international expansion by examining the entry and survival of foreign subsidiaries in the U.S. computer and pharmaceutical industries over the 1974-89 period. Using a hazard rate model, we examine the effects of (1) diversification strategies, (2) entry strategies, and (3) organizational learning and experience on the survival probabilities of foreign subsidiaries. The results show a higher exit rate for foreign acquisitions and joint ventures than for subsidiaries established through greenfield investments. The results also indicate a higher exit rate for subsidiaries that diversify than for those that stay in the parent firm's main product areas. Finally, the results show that firms benefit from learning and experience in foreign operations, which improves the chances of success for subsequent foreign investments. These findings shed light on the dynamic process of international expansion and the evolution of the multinational corporation.

Extending the Chain of Relationships Among Organizational Justice, Social Exchange, and Employee Reactions: The Role of Contract Violations

Abstract: This study extends prior research on the chain of relationships among organizational justice, social exchange relationships, and employee reactions by investigating the proposed mediating role of psychological contract violations. Results obtained from a longitudinal design examining a sample of 191 employees provide strong support for the proposal, enhance support for chain directionality, validate theoretical predictions about determinants of contract violations, and provide initial evidence supporting a proposed integration of perceived organizational support and psychological contract theory.

Cluster-Based Cooperative Spectrum Sensing in Cognitive Radio Systems

Abstract: In cognitive radio systems, secondary users can be coordinated to perform cooperative spectrum sensing so as to detect the primary user more accurately. However, when the sensing observations are forwarded to a common receiver through fading channels, the sensing performance can be severely degraded. In this paper, we propose a cluster-based cooperative spectrum sensing method to improve the sensing performance. By separating all the secondary users into a few clusters and selecting the most favorable user in each cluster to report to the common receiver, the proposed method can exploit the user selection diversity so that the sensing performance can be enhanced. Furthermore, decision fusion and energy fusion are both studied and the analytical performance results are given. Numerical results show that the sensing performance is improved significantly as opposed to conventional spectrum sensing.

Is There an "Academic Vocabulary"?

Abstract: This article considers the notion of academic vocabulary: the assumption that students of English for academic purposes (EAP) should study a core of high frequency words because they are common in an English academic register. We examine the value of the term by using Coxhead's (2000) Academic Word List (AWL) to explore the distribution of its 570 word families in a corpus of 3.3 million words from a range of academic disciplines and genres. The findings suggest that although the AWL covers 10.6% of the corpus, individual lexical items on the list often occur and behave in different ways across disciplines in terms of range, frequency, collocation, and meaning. This result suggests that the AWL might not be as general as it was intended to be and, more importantly, questions the widely held assumption that students need a single core vocabulary for academic study. We argue that the different practices and discourses of disciplinary communities undermine the usefulness of such lists and recommend that teachers help students develop a more restricted, discipline-based lexical repertoire.

Intergenerational Trade, Longevity, and Economic Growth

Abstract: We develop an overlapping-generations model of endogenous growth in which human capital is the engine of growth and the generations are linked through material and emotional interdependencies within the family. Parents invest in their children to achieve both old-age support (care) and emotional gratification, and material support from children is determined through self-enforcing implicit contracts. We show that optimal intergenerational trade can then lead to maximization of growth opportunities. Our model produces a theory of the demographic transition linking longevity, fertility, and economic growth. We also show that while population aging may raise the growth rate, an increase in young- ge longevity is likely to produce a greater increase in the growth rate and a reduction in the fertility rate in a growth equilibrium. These predictions and the model's implications concerning the behavior of private savings during the takeoff period appear consistent with empirical evidence.

Loyalty to Supervisor vs. Organizational Commitment: Relationships to Employee Performance in China

Abstract: In this research, we investigated the relationship between loyalty to supervisor and employee's in-role and extra-role performance in comparison with that of organizational commitment in the People's Republic of China. Two studies were conducted. In the first study, a five-dimension loyalty to supervisor scale was developed and validated. In the second study, the relationships between loyalty to supervisor, organizational commitment and employee performance were examined. Results indicated that loyalty to supervisor was more strongly associated with both in-role and extra-role performance than organizational commitment. The findings are discussed in terms of their implications for future research and management practices in cross-cultural settings.

Cooperative Spectrum Sensing for Cognitive Radios under Bandwidth Constraints

Abstract: In cognitive radio systems, cooperative spectrum sensing is conducted among the cognitive users so as to detect the primary user accurately. However, when the number of cognitive users tends to be very large, the bandwidth for reporting their sensing results to the common receiver will be very huge. In this paper, the authors employ a censoring method with quantization to decrease the average number of sensing bits to the common receiver. By censoring the collected local observations, only the users with enough information will send their local one bit decisions (0 or 1) to the common receiver. The performance of spectrum sensing is investigated for both perfect and imperfect reporting channels. Numerical results show that the average number of sensing bits decreases greatly at the expense of a little sensing performance loss.