Computer Sciences Commons^™

Two-Sided Exact Tests And Matching Confidence Intervals For Discrete Data, Michael P. Fay

The R Journal

There is an inherent relationship between two-sided hypothesis tests and confidence intervals. A series of two-sided hypothesis tests may be inverted to obtain the matching 100(1-)% confidence interval defined as the smallest interval that contains all point null parameter values that would not be rejected at the α level. Unfortunately, for discrete data there are several different ways of defining two-sided exact tests and the most commonly used two sided exact tests are defined one way, while the most commonly used exact confidence intervals are inversions of tests defined another way. This can lead to inconsistencies where the exact test …

Neuralnet: Training Of Neural Networks, Franke Günther, Stefan Fritsch

The R Journal

Artificial neural networks are applied in many situations. neuralnet is built to train multi-layer perceptrons in the context of regression analyses, i.e. to approximate functional relationships between covariates and response variables. Thus, neural networks are used as extensions of generalized linear models. neuralnet is a very flexible package. The back propagation algorithm and three versions of resilient back-propagation are implemented and it provides a custom-choice of activation and error function. An arbitrary number of covariates and response variables as well as of hidden layers can theoretically be included. The paper gives a brief introduction to multi-layer perceptrons and resilient back-propagation …

Mcmc For Generalized Linear Mixed Models With Glmmbugs, Patrick Brown, Lutong Zhou

The R Journal

The glmmBUGS package is a bridging tool between Generalized Linear Mixed Models (GLMMs) in R and the BUGS language. It provides a simple way of performing Bayesian inference using Markov Chain Monte Carlo (MCMC) methods, taking a model formula and data frame in R and writing a BUGS model file, data file, and initial values files. Functions are provided to reformat and summarize the BUGS results. A key aim of the package is to provide files and objects that can be modified prior to calling BUGS, giving users a platform for customizing and extending the models to accommodate a wide …

Glmperm: A Permutation Of Regressor Residuals Test For Inference In Generalized Linear Models, Wiebke Werft, Axel Benner

The R Journal

We introduce a new R package called glmperm for inference in generalized linear models especially for small and moderate-sized data sets. The inference is based on the per mutation of regressor residuals test introduced by Potter (2005). The implementation of glmperm outperforms currently available permutation test software as glmperm can be applied in situations where more than one covariate is involved.

Tmvtnorm: A Package For The Truncated Multivariate Normal Distribution, Stefan Wilhelm, B. G. Manjunath

The R Journal

In this article we present tmvtnorm, an R package implementation for the truncated multivariate normal distribution. We consider random number generation with rejection and Gibbs sampling, computation of marginal densities as well as computation of the mean and co variance of the truncated variables. This contribution brings together latest research in this field and provides useful methods for both scholars and practitioners when working with truncated normal variables.

The Google Search Engine, Ganimete Perçuku

Theses and Dissertations

Përgjithësisht makina kërkuese Google paraqitet si sistemi i kompjuterëve të projektuar për kërkimin e informatave në ueb. Google mundohet t’i kuptojë kërkesat e njerëzve në mënyrë “njerëzore”, dhe t’iu kthej atyre përgjigjen në formën të qartë. Por, ky synim nuk është as afër ideales dhe realizimi i tij sa vjen e vështirësohet me zgjerimin eksponencial që sot po përjeton ueb-i. Google, paraqitet duke ngërthyer në vetvete shqyrtimin e pjesëve që e përbëjnë, atyre në të cilat sistemi mbështetet, dhe rrethinave tjera që i mundësojnë sistemit të funksionojë pa probleme apo të përtërihet lehtë nga ndonjë dështim eventual. Procesi i grumbullimit …

Max Flow Spill Code Placement Algorithm Implemented In Gcc 4.4.3, Stephen Robert Beard

Computer Engineering

The placement of spill code plays an important role in the register allocator of an optimizing compiler. Many computer architectures possess a register linkage convention that dictates which registers are preserved across function calls and which are not. This project addresses the problem of optimizing spill code that is associated with register linkage conventions.

This algorithm was created by Dr. Chris Lupo and is described in the paper Beyond Register Allocation: a Novel Algorithm for Spill-Code Placement. The algorithm was implemented for GCC 2.5.7 for a PA-RISC architecture [4]. The work in this project will involve porting the existing code …

Wii-Mote Head Tracking: A Three Dimensional Virtual Reality Display, David Fairman

Computer Engineering

The goal of this project is to create a customizable three dimensional virtual reality display on a system available to any non-technical user. This System will use the infrared camera component of a standard Nintendo Wii-mote to track a user's head motions in all six major directions. The virtual reality will be a customizable image projected onto a screen or simply shown on a computer or TV monitor. In order to appear 3-dimensional, the image will continually change according to the position of the user's head. As the user moves their head to the left and right, portions of the …

Janus: From Workflows To Semantic Provenance And Linked Open Data, Paolo Missier, Satya S. Sahoo, Jun Zhao, Carole Goble, Amit P. Sheth

Kno.e.sis Publications

Data provenance graphs are form of metadata that can be used to establish a variety of properties of data products that undergo sequences of transformations, typically specified as workflows. Their usefulness for answering user provenance queries is limited, however, unless the graphs are enhanced with domain-specific annotations. In this paper we propose a model and architecture for semantic, domain-aware provenance, and demonstrate its usefulness in answering typical user queries. Furthermore, we discuss the additional benefits and the technical implications of publishing provenance graphs as a form of Linked Data. A prototype implementation of the model is available for data produced …

Semantic Context Modeling With Maximal Margin Conditional Random Fields For Automatic Image Annotation, Yu Xiang, Xiangdong Zhou, Zuotao Liu, Tat-Seng Chua, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

Context modeling for Vision Recognition and Automatic Image Annotation (AIA) has attracted increasing attentions in recent years. For various contextual information and resources, semantic context has been exploited in AIA and brings promising results. However, previous works either casted the problem into structural classification or adopted multi-layer modeling, which suffer from the problems of scalability or model efficiency. In this paper, we propose a novel discriminative Conditional Random Field (CRF) model for semantic context modeling in AIA, which is built over semantic concepts and treats an image as a whole observation without segmentation. Our model captures the interactions between semantic …

Prediction Of Protein Subcellular Localization: A Machine Learning Approach, Kyong Jin Shim

Research Collection School Of Computing and Information Systems

Subcellular localization is a key functional characteristic of proteins. Optimally combining available information is one of the key challenges in today's knowledge-based subcellular localization prediction approaches. This study explores machine learning approaches for the prediction of protein subcellular localization that use resources concerning Gene Ontology and secondary structures. Using the spectrum kernel for feature representation of amino acid sequences and secondary structures, we explore an SVM-based learning method that classifies six subcellular localization sites: endoplasmic reticulum, extracellular, Golgi, membrane, mitochondria, and nucleus.

Employee Time Scheduling, Mark Peter Smith

Computer Science and Software Engineering

Small business managers face the common problem of employee time scheduling. There is a solution to this problem in the form of an application called Lemming Scheduler. Lemming Scheduler is a Java based employee time scheduling program. Its features include a desktop based application that stores employee and business information as well as a web interface for employees to view schedules and update availability. The desktop application uses employee and shift information to automatically generate schedules. The generated schedules are viewable by employees outside of work by way of the web interface. Lemming Scheduler provides a light weight interface for …

Information Hiding Using Stochastic Diffusion For The Covert Transmission Of Encrypted Images, Jonathan Blackledge

Conference papers

A principal weakness of all encryption systems is that the output data can be `seen' to be encrypted. In other words, encrypted data provides a 'flag' on the potential value of the information that has been encrypted. In this paper, we provide a novel approach to `hiding' encrypted data in a digital image. We consider an approach in which a plaintext image is encrypted with a cipher using the processes of `stochastic diffusion' and the output quantized into a 1-bit array generating a binary image cipher-text. This output is then `embedded' in a host image which is undertaken either in …

Re-Solving Stochastic Programming Models For Airline Revenue Management, Lijian Chen, Tito Homem-De-Mello

MIS/OM/DS Faculty Publications

We study some mathematical programming formulations for the origin-destination model in airline revenue management. In particular, we focus on the traditional probabilistic model proposed in the literature. The approach we study consists of solving a sequence of two-stage stochastic programs with simple recourse, which can be viewed as an approximation to a multi-stage stochastic programming formulation to the seat allocation problem. Our theoretical results show that the proposed approximation is robust, in the sense that solving more successive two-stage programs can never worsen the expected revenue obtained with the corresponding allocation policy. Although intuitive, such a property is known not …

Analysis And Transformation Of Pipe-Like Web Mashups For End User Programmers, Kathryn T. Stolee

School of Computing: Dissertations, Theses, and Student Research

Mashups are becoming increasingly popular as end users are able to easily access, manipulate, and compose data from several web sources. To support end users, communities are forming around mashup development environments that facilitate sharing code and knowledge. We have observed, however, that end user mashups tend to suffer from several deficiencies, such as inoperable components or references to invalid data sources, and that those deficiencies are often propagated through the rampant reuse in these end user communities.

In this work, we identify and specify ten code smells indicative of deficiencies we observed in a sample of 8,051 pipe-like web …

Optimization Of P3ht-Pcbm Polymer Solar Cells Through Device Simulation And Manufacturing, James Boom

Computer Engineering

Given a good model and implementation of that model, computer simulation can be used to reduce the time and material costs of research. To this end I worked with other students to manufacture, test and simulate the single layer P3HT-PCBM solar cell. Using the data collected from this project, future work can then be done with the project's simulator to further optimize these types of solar cell devices.

Provenance Management In Parasite Research, Vinh Nguyen, Priti Parikh, Satya S. Sahoo, Amit P. Sheth

Kno.e.sis Publications

The objective of this research is to create a semantic problem solving environment (PSE) for human parasite Trypanosoma cruzi. As a part of the PSE, we are trying to manage provenance of the experiment data as it is generated. It requires to capture the provenance which is often collected through web forms used by biologists to input the information about experiments they conduct. We have created Parasite Experiment Ontology (PEO) that represents provenance information used in the project. We have modified the back end which processes the data gathered from biologists, generates RDF triples and serializes them into the triple …

Predictive Yasir: High Security With Lower Latency In Legacy Scada, Rouslan V. Solomakhin

Dartmouth College Master’s Theses

Message authentication with low latency is necessary to ensure secure operations in legacy industrial control networks, such as power grid networks. Previous authentication solutions by our lab and others looked at single messages and incurred noticeable latency. To reduce this latency, we develop Predictive YASIR, a bump-in-the-wire device that looks at broader patterns of messages. The device (1) predicts the incoming plaintext based on previous observations; (2) compresses, encrypts, and authenticates data online; and (3) pre-sends a part of ciphertext before receiving the whole plaintext. I demonstrate the performance properties of this approach by implementing it in the Scalable Simulation …

Reducing Cluster Power Consumption By Dynamically Suspending Idle Nodes, Brian Michael Oppenheim

Master's Theses

Close to 1% of the world's electricity is consumed by computer servers. Given that the increased use of electricity raises costs and damages the environment, optimizing the world's computing infrastructure for power consumption is worthwhile. This thesis is one attempt at such an optimization. In particular, I began by building a cluster of 6 Intel Atom based low-power nodes to perform work analogous to data center clusters. Then, I installed a version of Hadoop modified with a novel power management system on the cluster. The power management system uses different algorithms to determine when to turn off idle nodes in …

Convergence Of The Sinc Method Applied To Volterra Integral Equations, M. Zarebnia, J. Rashidinia

Applications and Applied Mathematics: An International Journal (AAM)

A collocation procedure is developed for the linear and nonlinear Volterra integral equations, using the globally defined Sinc and auxiliary basis functions. We analytically show the exponential convergence of the Sinc collocation method for approximate solution of Volterra integral equations. Numerical examples are included to confirm applicability and justify rapid convergence of our method.

The Teaching Of Equation Solving: Approaches In Standards-Based And Traditional Curricula In The United States, Jinfa Cai, Bikai Nie, John Moyer

Mathematics, Statistics and Computer Science Faculty Research and Publications

This paper discusses the approaches to teaching linear equation solving that are embedded in a Standards-based mathematics curriculum (Connected Mathematics Program or CMP) and in a traditional mathematics curriculum (Glencoe Mathematics) in the United States. Overall, the CMP curriculum takes a functional approach to teaching equation solving, while Glencoe Mathematics takes a structural approach. The functional approach emphasizes the important ideas of change and variation in situations and contexts. It also emphasizes the representation of relationships between variables. The structural approach, on the other hand, requires students to work abstractly with symbols and follow procedures in a systematic way. …

Measuring Atmospheric Scattering From Digital Images Of Urban Scenery Using Temporal Polarization-Based Vision, Tarek El-Gaaly

Archived Theses and Dissertations

Suspended atmospheric particles (particulate matter) are a form of air pollution that visually degrades urban scenery and is hazardous to human health and the environment. Current environmental monitoring devices are limited in their capability of measuring average particulate matter (PM) over large areas. Quantifying the visual effects of haze in digital images of urban scenery and correlating these effects to PM levels is a vital step in more practically monitoring our environment. Current image haze extraction algorithms remove all the haze from the scene and hence produce unnatural scenes for the sole purpose of enhancing vision. We present two algorithms …

An Extended Configurable Uml Activity Diagram And A Transformation Algorithm For Business Process Reference Modeling, Yosra Osama Badr

Archived Theses and Dissertations

Enterprise Resource Planning (ERP) solutions provide generic off-the-shelf reference models usually known as "best practices". The configuration !individualization of the reference model to meet specific requirements of business end users however, is a difficult task. The available modeling languages do not provide a complete configurable language that could be used to model configurable reference models. More specifically, there is no algorithm that monitors the transformation of configurable UML Activity Diagram (AD) models while preserving the syntactic correctness of the model. To fill these gaps we propose an extended UML AD modeling language which we named Configurable UML Activity Diagram (C-UML …

Information Risk And Underwriter Switching In Seos: Evidence From China, Wei Luo, Pingui Rao, Heng Yue

Research Collection School Of Accountancy

In this paper we examine whether information risk affects underwriter switching in a seasoned equity offering (SEO) process. Building on previous research, we hypothesize that SEO firms and underwriters associate with one another by mutual choice, and firms with a low degree of information risk tend to match up with prestigious underwriters. Using a sample of SEO firms in China and employing accruals quality as a proxy of information risk, we find evidence consistent with our hypothesis: the information risk and the initial public offering (IPO) underwriters’ reputation at the time of the SEO jointly determine the probability that the …

Real-Time Visualizations Of Ocean Data Collected By The Norus Glider, Daniel M. Medina

Master's Theses

Scientific visualization computer applications generate visual representations of large and complex sets of science data. These types of applications allow scientists to gain greater knowledge and insight into their data. For example, the visualization of environmental data is of particular interest to biologists when trying to understand how complex variables interact. Modern robotics and sensors have expanded the ability to collect environmental data, thus, the size and variety of these data-sets have likewise grown. Oftentimes, the collected data are deposited into files and databases where they sit in their separate and unique formats. Without easy to use visualization tools, it …

Using Hadoop And Cassandra For Taxi Data Analytics: A Feasibility Study, Alvin Jun Yong Koh, Xuan Khoa Nguyen, C. Jason Woodard

Research Collection School Of Computing and Information Systems

This paper reports on a preliminary study to assess the feasibility of using the Open Cirrus Cloud Computing Research testbed to provide offline and online analytical support for taxi fleet operations. In the study, we benchmarked the performance gains from distributing the offline analysis of GPS location traces over multiple virtual machines using the Apache Hadoop implementation of the MapReduce paradigm. We also explored the use of the Apache Cassandra distributed database system for online retrieval of vehicle trace data. While configuring the testbed infrastructure was straightforward, we encountered severe I/O bottlenecks in running the benchmarks due to the lack …

A Hybrid Method To Detect Deflation Fraud In Cost-Per-Action Online Advertising, Xuhua Ding

Research Collection School Of Computing and Information Systems

Web advertisers prefer the cost-per-action (CPA) advertisement model whereby an advertiser pays a web publisher according to the actual amount of transactions, rather than the volume of advertisement clicks. The main obstacle for a wide deployment of this model is the deflation fraud. Namely, a dishonest advertiser under-reports the transaction count in order to discharge less. In this paper, we present a mechanism to detect such a fraud using a hybrid of cryptography and probability tools. With the assistance from a small number of users, the publisher can detect deflation fraud with a success probability growing exponentially with the fraud …

Income, Endogenous Market Structure And Innovation, Mei Lin, Shaojin Li, Andrew B. Whinston

Research Collection School Of Computing and Information Systems

We investigate the effect of income distribution on R&D in a dynamic framework. Our model captures both the infinite R&D race among heterogeneous innovators and a market where successful innovators generate revenues. The market structure of successful innovations is endogenous – firms produce vertically differentiated substitute goods and compete in price. Based on firms' equilibrium market revenues, we derive numerical solutions of the Markov perfect equilibrium innovation rate of the dynamic problem. A key insight in our results is that explicitly modeling price competition and the market structure plays an important role in evaluating the impact of rising income inequality …

Anytime Planning For Decentralized Pomdps Using Expectation Maximization, Akshat Kumar, Shlomo Zilberstein

Research Collection School Of Computing and Information Systems

Decentralized POMDPs provide an expressive framework for multi-agent sequential decision making. While finite-horizon DECPOMDPs have enjoyed signifcant success, progress remains slow for the infinite-horizon case mainly due to the inherent complexity of optimizing stochastic controllers representing agent policies. We present a promising new class of algorithms for the infinite-horizon case, which recasts the optimization problem as inference in a mixture of DBNs. An attractive feature of this approach is the straightforward adoption of existing inference techniques in DBNs for solving DEC-POMDPs and supporting richer representations such as factored or continuous states and actions. We also derive the Expectation Maximization (EM) …

Efficient Mutual Nearest Neighbor Query Processing For Moving Object Trajectories, Yunjun Gao, Baihua Zheng, Gencai Chen, Qing Li, Chun Chen, Gang Chen

Research Collection School Of Computing and Information Systems

Given a set D of trajectories, a query object q, and a query time extent Γ, a mutual (i.e., symmetric) nearest neighbor (MNN) query over trajectories finds from D, the set of trajectories that are among the k1 nearest neighbors (NNs) of q within Γ, and meanwhile, have q as one of their k2 NNs. This type of queries is useful in many applications such as decision making, data mining, and pattern recognition, as it considers both the proximity of the trajectories to q and the proximity of q to the trajectories. In this paper, we first formalize MNN search …