Digital Commons Network^™

Robust Classification Of Event-Related Potential For Brain-Computer Interface, Manoj Thulasidas

Research Collection School Of Computing and Information Systems

We report the implementation of a text input application (speller) based on the P300 event related potential. We obtain high accuracies by using an SVM classifier and a novel feature. These techniques enable us to maintain fast performance without sacrificing the accuracy, thus making the speller usable in an online mode. In order to further improve the usability, we perform various studies on the data with a view to minimizing the training time required. We present data collected from nine healthy subjects, along with the high accuracies (of the order of 95% or more) measured online. We show that the …

A Qualitative Investigation On Consumer Trust In Mobile Commerce, Keng Siau, Hong Sheng, Fiona Fui-Hoon Nah, Sidney A. Davis

Research Collection School Of Computing and Information Systems

Mobile commerce represents a significant development in e-commerce, offering accessibility, ubiquity, mobility, and localisation to users. Despite the potential of mobile commerce, trust is a major obstacle in its adoption and development. Many consumers feel uncomfortable with the idea of conducting commerce over wireless, hand-held devices. The focus of this research is to understand trust in mobile commerce and to identify factors that are important for trust development. The research builds on Siau and Shen's framework which depicts two key factors influencing trust in mobile commerce. This research not only validates and expands on the existing framework, but also provides …

A Spectroscopy Of Texts For Effective Clustering, Wenyuan Li, Wee-Keong Ng, Kok-Leong Ong, Ee Peng Lim

Research Collection School Of Computing and Information Systems

For many clustering algorithms, such as k-means, EM, and CLOPE, there is usually a requirement to set some parameters. Often, these parameters directly or indirectly control the number of clusters to return. In the presence of different data characteristics and analysis contexts, it is often difficult for the user to estimate the number of clusters in the data set. This is especially true in text collections such as Web documents, images or biological data. The fundamental question this paper addresses is: ldquoHow can we effectively estimate the natural number of clusters in a given text collection?rdquo. We propose to use …

Sclope: An Algorithm For Clustering Data Streams Of Categorical Attributes, Kok-Leong Ong, Wenyuan Li, Wee-Keong Ng, Ee Peng Lim

Research Collection School Of Computing and Information Systems

Clustering is a difficult problem especially when we consider the task in the context of a data stream of categorical attributes. In this paper, we propose SCLOPE, a novel algorithm based on CLOPE's intuitive observation about cluster histograms. Unlike CLOPE however, our algorithm is very fast and operates within the constraints of a data stream environment. In particular, we designed SCLOPE according to the recent CluStream framework. Our evaluation of SCLOPE shows very promising results. It consistently outperforms CLOPE in speed and scalability tests on our data sets while maintaining high cluster purity; it also supports cluster analysis that other …

Shared-Storage Auction Ensures Data Availability, Hady W. Lauw, Siu-Cheung Hui, Edmund M. K. Lai

Research Collection School Of Computing and Information Systems

Most current e-auction systems are based on the client-server architecture. Such centralized systems provide a single point of failure and control. In contrast, peer-to-peer systems permit distributed control and minimize individual node and link failures' impact on the system. The shared-storage-based auction model described decentralizes services among peers to share the required processing load and aggregates peers' resources for common use. The model is based on the principles of local computation at each peer, direct inter-peer communication, and a shared storage space.

Proxy Signature Scheme With Multiple Original Signers For Wireless E-Commerce Applications, Guilin Wang, Feng Bao, Jianying Zhou, Robert H. Deng

Research Collection School Of Computing and Information Systems

In a proxy signature scheme, a user delegates his/her signing capability to another user in such a way that the latter can sign messages on behalf of the former. We propose an efficient and secure proxy signature scheme with multiple original signers. Our scheme is suitable for wireless electronic commerce applications, since the overheads of computation and communication are low. As an example, we present an electronic air ticket booking scheme for wireless customers.

Singular Value Decomposition Of Analogs Of Gbr 12909, Anna Fiorentino

Theses

Analogs of GBR 12909 are drugs that could potentially be used to treat cocaine addiction. Singular Value Decomposition (SVD) is a multivariate analysis technique used to show relationships between the data and the variables associated with the data. The input data consists of the conformers of each analog (DM324, 728 conformers; TP250, 739 conformers) along with the eight torsional angles (Al, A2, B1-B6). A novel scaling technique was developed to address the problem of data circularity by subtracting the values of the torsional angles of the global energy minimum conformation from those of each conformer.

In SVD the original data …

Slimsvm : A Simple Implementation Of Support Vector Machine For Analysis Of Microarray Data, Avik Karmaker

Theses

Support Vector Machine (SVM) is a supervised machine learning technique being widely used in multiple areas of biological analysis including microarray data analysis. SlimSVM has been developed with the intention of replacing OSU SVM as the classification component of GenoIterSVM in order to make it independent of other SVM packages. GenolterSVM, developed by Dr. Marc Ma, is a SVM implementation with an iterative refinement algorithm for improved accuracy of classification of genotype microarray data. SlimSVM is an object-oriented, modular, and easy-to-use implementation written in C++. It supports dot (linear) and polynomial (non-linear) kernels. The program has been tested with artificial …

An Ecological Display For Robot Teleoperation, Robert W. Ricks

Theses and Dissertations

This thesis presents an interface for effectively teleoperating robots that combines an ecological display of range and video information with quickening based on dead-reckoning prediction. This display is built by viewing range and video information from a virtual camera trailing the robot. This is rendered in 3-D by using standard hardware acceleration and 3-D graphics software. Our studies demonstrate that this interface improves performance for most people, including those that do not have much previous experience with robotics. These studies involved 32 test subjects in a simulated environment and 8 in the real world. Subjects were required to drive the …

Sb9: The Ninth Catalogue Of Spectroscopic Binary Orbits, Dimitri Pourbaix, Andrei A. Tokovinin, Alan H. Batten, Francis C. Fekel, William I. Hartkopf, Hugo Levato, Nidia I. Morrell, Guillermo Torres, Stephane Udry

Information Systems and Engineering Management Research Publications

The Ninth Catalogue of Spectroscopic Binary Orbits (http://sb9.astro.ulb.ac.be) continues the series of compilations of spectroscopic orbits carried out over the past 35 years by Batten and collaborators. As of 2004 May 1st, the new Catalogue holds orbits for 2386 systems. Some essential differences between this catalogue and its predecessors are outlined and three straightforward applications are presented: (1) completeness assessment: period distribution of SB1s and SB2s; (2) shortest periods across the H-R diagram; (3) period-eccentricity relation.

Direct Least-Squares Ellipse Fitting, Jane Courtney, Annraoi Depaor

Conference Papers

Many biological and astronomical forms can be best represented by ellipses. While some more complex curves might represent the shape more accurately, ellipses have the advantage that they are easily parameterised and define the location, orientation and dimensions of the data more clearly. In this paper, we present a method of direct least-squares ellipse fitting by solving a generalised eigensystem. This is more efficient and more accurate than many alternative approaches to the ellipse-fitting problem such as fuzzy c-shells clustering and Hough transforms. This method was developed for human body modelling as part of a larger project to design a …

Improving Throughput By Link Distance Control In A Multi-Rate Ad Hoc Network, Xiaoxin Wu, Gang Ding, Bharat Bhargava, Shan Lei

Department of Computer Science Technical Reports

No abstract provided.

Ica-Fx Features For Classification Of Singing Voice And Instrumental Sound, Tat-Wan Leung, Chong-Wah Ngo, Rynson W. H Lau

Research Collection School Of Computing and Information Systems

This paper describes a new approach in locating the segments of singing voice in pop musical songs. Initially, GLR distance measure is employed to temporally detect the boundaries of singing voices and instrumental sounds. ICAFX is then adopted to extract the independent components of acoustic features for SVM classification. Experimental results indicate that ICA-FX can improve the classification performance by significantly reducing the independent components that are not related to class label information.

Graph Based Image Matching, Hui Jiang, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

Given two or more images, we can define different but related problems on pattern matching such as image registration, pattern detection and localization, and common pattern discovery. These problems have different levels of purpose and difficulties, as a result, often associate with different solutions. In this paper, we propose a novel approach to solve these problems under a unified framework based on graph matching. We first split the images into small blocks and represent each block as a node in a bipartite graph. A maximum weighted bipartite graph matching algorithm is then employed in an iterative way to find the …

Novel Seed Selection For Multiple Objects Detection And Tracking, Zailiang Pan, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

This paper proposes a unified approach for initializing, detecting and tracking of multiple moving objects. Object initialization is achieved through novel seed selection which is adaptively activated, depending on the quality of tracking, to select the best possible frames along the temporal direction for object detection. EM algorithm is then employed to robustly segment and detect multiple objects in a selected frame. Each detected object is represented by an appearance-based model and mean shift tracking procedure is adopted to rapidly and effectively track the target objects.

Gesture Tracking And Recognition For Lecture Video Editing, Feng Wang, Chong-Wah Ngo, Ting-Chuen Pong

Research Collection School Of Computing and Information Systems

This paper presents a gesture based driven approach for video editing. Given a lecture video, we adopt novel approaches to automatically detect and synchronize its content with electronic slides. The gestures in each synchronized topic (or shot) are then tracked and recognized continuously. By registering shots and slides and recovering their transformation, the regions where the gestures take place can be known. Based on the recognized gestures and their registered positions, the information in slides can be seamlessly extracted, not only to assist video editing, but also to enhance the quality of original lecture video.

Deformable Geometry Model Matching By Topological And Geometric Signatures, Kwok-Leung Tam, Rynson W.H. Law, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

In this paper, we present a novel method for efficient 3D model comparison. The method matches highly deformed models by comparing topological and geometric features. First, we propose “Bi-directional LSD analysis” to locate reliable topological points and rings. Second, based on these points and rings, a set of bounded regions are extracted as topological features. Third, for each bounded region, we capture additional spatial location, curvature and area distribution as geometric data. Fourth, to model the topological importance of each bounded region, we capture its effective area as weight. By using “Earth Mover Distance” as a distance measure between two …

Investigating Deception In Cyberspace, Holtjona Galanxhi, Fiona Fui-Hoon Nah

Research Collection School Of Computing and Information Systems

The use of anthropomorphic avatars provides Internet users the opportunity and freedom to present their desired identity. As cyberspace becomes a heaven for deceptive behavior, human-computer interaction research will need to be carried out to study and understand these deceptive behaviors. The objective of this research is to investigate the behavior of deceivers and truthtellers in the cyberspace environment and to examine how their intention to deceive may influence the choice of their avatars. The effect of using avatars to support communication will also be studied. This research provides insights on communication and trust issues in avatar environments.

Packet-Loss Resilient Coding Scheme With Only Xor Operations, Gui Liang Feng, Robert H. Deng, Feng Bao

Research Collection School Of Computing and Information Systems

Application of forward error correction to recover lost packets in higher layers of communication networks is receiving increasing attention. Most of the previous proposals for packet loss recovery use symbol-oriented Reed-Solomon codes operating in symbol erasure-correction mode. A Reed-Solomon code is optimal in the sense that it is maximal distance separable; however, the decoding speed of a Reed-Solomon code is slow since it involves operations over GF(2m) using lookup tables. A packet-oriented (n, k)/(m, l) packet-loss resilient code based on an (n, k) Reed-Solomon code over GF(2m) is given. The code accepts k-packet information sequences and encodes them into n-packet …

Exploiting Information Theory For Adaptive Mobility And Resource Management In Future Cellular Networks, Abhishek Roy, Sajal K. Das, Archan Misra

Research Collection School Of Computing and Information Systems

We utilize tools from information theory to develop adaptive algorithms for two key problems in cellular networks: location tracking and resource management. The use of information theory is motivated by the fundamental observation that overheads in many aspects of mobile computing can be traced to the randomness or uncertainty in an individual user's movement behavior. We present a model-independent information-theoretic approach for estimating and managing this uncertainty, and relate it to the entropy or information content of the user's movement process. Information-theoretic mobility management algorithms are very simple, yet reduce overhead by ∼80 percent in simulated scenarios by optimally adapting …

Group-Based Relevance Feedback With Support Vector Machine Ensembles, Steven C. H. Hoi, Michael R. Lyu

Research Collection School Of Computing and Information Systems

Support vector machines (SVMs) have become one of the most promising techniques for relevance feedback in content-based image retrieval (CBIR). Typical SVM-based relevance feedback techniquessimply apply the strict binary classifications: positive (relevant) class and negative (irrelevant) class. However, in a real-world relevance feedback task, it is more reasonable and practical to assume the data come from multiple positive classes and one negative class. In order to formulate an effective relevance feedback algorithm, we propose a novel group-based relevance feedback scheme constructed with the SVM ensembles technique. Experiments are conducted to evaluate the performance of our proposed scheme and the traditional …

Pre-Conference Proceedings Of The Focus Symposium On Intelligent Software Systems For The New Infostructure: Baden-Baden, Germany, Collaborative Agent Design Research Center

Collaborative Agent Design (CAD) Research Center

Over the past several years the papers in this annual series of symposia have increasingly centered on the realization of a human-computer collaboration environment in which computer-based software agents with reasoning capabilities provide meaningful support to human decision makers. It is therefore quite appropriate that the first paper in the 2004 Proceedings should address the historical evolutionary path of ‘intelligent’ software leading to the goal of a semantic Web environment. The realization of this goal is now in sight, driven by public security threats that are increasingly relying on technology for effective countermeasures.

The Evolution Of Intelligent Computer Software And The Semantic Web, Jens G. Pohl

Collaborative Agent Design (CAD) Research Center

The purpose of this paper is to trace the evolution of intelligent software from data-centric applications that essentially encapsulate their data environment to ontology-based applications with automated reasoning capabilities. The author draws a distinction between human intelligence and component capabilities within a more general definition of intelligence, which may be embedded in computer software. The primary vehicle in the quest for intelligent software has been the gradual recognition of the central role played by data and information, rather than the logic and functionality of the application. The three milestones in this evolution have been: the separation of data management from …

Feature Weighting Using Neural Networks, Tony R. Martinez, Xinchuan Zeng

Faculty Publications

In this work we propose a feature weighting method for classification tasks by extracting relevant information from a trained neural network. This method weights an attribute based on strengths (weights) of related links in the neural network, in which an important feature is typically connected to strong links and has more impact on the outputs. This method is applied to feature weighting br the nearest neighbor classifier and is tested on 15 real-world classification tasks. The results show that it can improve the nearest neighbor classifier on 14 of the 15 tested tasks, and also outperforms the neural network on …

Softprop: Softmax Neural Network Backpropagation Learning, Tony R. Martinez, Michael E. Rimer

Faculty Publications

Multi-layer backpropagation, like many learning algorithms that can create complex decision surfaces, is prone to overfitting. Softprop is a novel learning approach presented here that is reminiscent of the softmax explore-exploit Q-learning search heuristic It fits the problem while delaying settling into error minima to achieve better generalization and more robust learning. This is accomplished by blending standard SSE optimization with lazy training, a new objective function well suited to learning classification tasks, to form a more stable learning model. Over several machine learning data sets, softprop reduces classification error by 17.1 percent and the variance in results by 38.6 …

Discovering Transcriptional Regulatory Rules From Gene Expression And Tf-Dna Binding Data By Decision Tree Learning, Jianhua Ruan, Weixiong Zhang

All Computer Science and Engineering Research

Background: One of the most promising but challenging task in the post-genomic era is to reconstruct the transcriptional regulatory networks. The goal is to reveal, for each gene that responds to a certain biological event, which transcription factors affect its transcription, and how several transcription factors coordinate to accomplish specific regulations. Results: Here we propose a supervised machine learning approach to address these questions. We build decision trees to associate the expression level of a gene with the transcription factor binding data of its promoter. From the decision trees, we extract regulatory rules that specify how the binding of a …

Enhancing Web Services Description And Discovery To Facilitate Composition, Preeda Rajasekaran, John A. Miller, Kunal Verma, Amit P. Sheth

Kno.e.sis Publications

Web services are in the midst of making the transition from being a promising technology to being widely used in the industry. However, most efforts to use Web services have been manual, thus slowing down the ever changing and dynamic businesses of today. In this paper, we contend that more expressive descriptions of Web services will lead to greater automation and thus provide more agility to businesses. We present the METEOR-S front-end tools for source code annotation and semantic Web service description generation. We also present WSDL-S, a language created for incorporating semantic descriptions in the industry wide accepted WSDL, …

Grouping Search-Engine Returned Citations For Person-Name Queries, Reema Al-Kamha

Theses and Dissertations

In this thesis we present a technique to group search-engine returned citations for person-name queries, such that the search-engine returned citations in each group belong to the same person. To group the returned citations we use a multi-faceted approach that considers evidence from three facets: (1) attributes, (2) links, and (3) page similarity. For each facet we generate a confidence matrix. Then we construct a final confidence matrix for all facets. Using a threshold, we apply a grouping algorithm on the final confidence matrix. The output is a group of search-engine returned citations, such that the citations in each group …

Probabilistic Disease Classification Of Expression-Dependent Proteomic Data From Mass Spectrometry Of Human Serum, Ryan H. Lilien, Hany Farid, Bruce R. Donald

Dartmouth Scholarship

We have developed an algorithm called Q5 for probabilistic classification of healthy vs. disease whole serum samples using mass spectrometry. The algorithm employs Principal Components Analysis (PCA) followed by Linear Discriminant Analysis (LDA) on whole spectrum Surface-Enhanced Laser Desorption/Ionization Time of Flight (SELDI-TOF) Mass Spectrometry (MS) data, and is demonstrated on four real datasets from complete, complex SELDI spectra of human blood serum.

Q5 is a closed-form, exact solution to the problem of classification of complete mass spectra of a complex protein mixture. Q5 employs a novel probabilistic classification algorithm built upon a dimension-reduced linear discriminant analysis. Our solution is …

Workflow Management Systems And Erp Systems: Differences, Commonalities, And Applications, Jorge Cardoso, Robert P. Bostrom, Amit P. Sheth

Kno.e.sis Publications

Two important classes of information systems, Workflow Management Systems(WfMSs) and Enterprise Resource Planning (ERP) systems, have been used to support e-business process redesign, integration, and management. While both technologies can help with business process automation, data transfer, and information sharing, the technological approach and features of solutions provided by WfMS and ERP are different. Currently, there is a lack of understanding of these two classes of information systems in the industry and academia, thus hindering their effective applications. In this paper, we present a comprehensive comparison between these two classes of systems. We discuss how the two types of systems …