Digital Commons Network^™

A Bayesian Decision Theoretical Approach To Supervised Learning, Selective Sampling, And Empirical Function Optimization, James Lamond Carroll

Theses and Dissertations

Many have used the principles of statistics and Bayesian decision theory to model specific learning problems. It is less common to see models of the processes of learning in general. One exception is the model of the supervised learning process known as the "Extended Bayesian Formalism" or EBF. This model is descriptive, in that it can describe and compare learning algorithms. Thus the EBF is capable of modeling both effective and ineffective learning algorithms. We extend the EBF to model un-supervised learning, semi-supervised learning, supervised learning, and empirical function optimization. We also generalize the utility model of the EBF to …

Quantification Of Artistic Style Through Sparse Coding Analysis In The Drawings Of Pieter Bruegel The Elder, James M. Hughes, Daniel J. Graham, Daniel N. Rockmore

Dartmouth Scholarship

Recently, statistical techniques have been used to assist art historians in the analysis of works of art. We present a novel technique for the quantification of artistic style that utilizes a sparse coding model. Originally developed in vision research, sparse coding models can be trained to represent any image space by maximizing the kurtosis of a representation of an arbitrarily selected image from that space. We apply such an analysis to successfully distinguish a set of authentic drawings by Pieter Bruegel the Elder from another set of well-known Bruegel imitations. We show that our approach, which involves a direct comparison …

Vowel Recognition From Continuous Articulatory Movements For Speaker-Dependent Applications, Jun Wang, Jordan R. Green, Ashok Samal, Tom D. Carrell

Department of Special Education and Communication Disorders: Faculty Publications

A novel approach was developed to recognize vowels from continuous tongue and lip movements. Vowels were classified based on movement patterns (rather than on derived articulatory features, e.g., lip opening) using a machine learning approach. Recognition accuracy on a single-speaker dataset was 94.02% with a very short latency. Recognition accuracy was better for high vowels than for low vowels. This finding parallels previous empirical findings on tongue movements during vowels. The recognition algorithm was then used to drive an articulation-to-acoustics synthesizer. The synthesizer recognizes vowels from continuous input stream of tongue and lip movements and plays the corresponding sound samples …

Prediction Of Brain Tumor Progression Using A Machine Learning Technique, Yuzhong Shen, Debrup Banerjee, Jiang Li, Adam Chandler, Yufei Shen, Frederic D. Mckenzie, Jihong Wang, Nico Karssemeijer (Ed.), Ronald M. Summers (Ed.)

Electrical & Computer Engineering Faculty Publications

A machine learning technique is presented for assessing brain tumor progression by exploring six patients' complete MRI records scanned during their visits in the past two years. There are ten MRI series, including diffusion tensor image (DTI), for each visit. After registering all series to the corresponding DTI scan at the first visit, annotated normal and tumor regions were overlaid. Intensity value of each pixel inside the annotated regions were then extracted across all of the ten MRI series to compose a 10 dimensional vector. Each feature vector falls into one of three categories:normal, tumor, and normal but progressed to …

A Boosting Framework For Visuality-Preserving Distance Metric Learning And Its Application To Medical Image Retrieval, Yang Liu, Rong Jin, Lily Mummert, Rahul Sukthankar, Adam Goode, Bin Zheng, Steven C. H. Hoi, Mahadev Satyanarayanan

Research Collection School Of Computing and Information Systems

Similarity measurement is a critical component in content-based image retrieval systems, and learning a good distance metric can significantly improve retrieval performance. However, despite extensive study, there are several major shortcomings with the existing approaches for distance metric learning that can significantly affect their application to medical image retrieval. In particular, "similarity" can mean very different things in image retrieval: resemblance in visual appearance (e.g., two images that look like one another) or similarity in semantic annotation (e.g., two images of tumors that look quite different yet are both malignant). Current approaches for distance metric learning typically address only one …

Predicting Flavonoid Ugt Regioselectivity With Graphical Residue Models And Machine Learning., Arthur Rhydon Jackson

Electronic Theses and Dissertations

Machine learning is applied to a challenging and biologically significant protein classification problem: the prediction of flavonoid UGT acceptor regioselectivity from primary protein sequence. Novel indices characterizing graphical models of protein residues are introduced. The indices are compared with existing amino acid indices and found to cluster residues appropriately. A variety of models employing the indices are then investigated by examining their performance when analyzed using nearest neighbor, support vector machine, and Bayesian neural network classifiers. Improvements over nearest neighbor classifications relying on standard alignment similarity scores are reported.

Noninvasive Estimation Of Pulmonary Artery Pressure Using Heart Sound Analysis, Aaron W. Dennis

Theses and Dissertations

Right-heart catheterization is the most accurate method for estimating pulmonary artery pressure (PAP). Because it is an invasive procedure it is expensive, exposes patients to the risk of infection, and is not suited for long-term monitoring situations. Medical researchers have shown that PAP influences the characteristics of heart sounds. This suggests that heart sound analysis is a potential noninvasive solution to the PAP estimation problem. This thesis describes the development of a prototype system, called PAPEr, which estimates PAP noninvasively using heart sound analysis. PAPEr uses patient data with machine learning algorithms to build models of how PAP affects heart …

Learning Condition-Specific Networks, Sushmita Roy

Computer Science ETDs

Condition-specific cellular networks are networks of genes and proteins that describe functional interactions among genes occurring under different environmental conditions. These networks provide a systems-level view of how the parts-list (genes and proteins) interact within the cell as it functions under changing environmental conditions and can provide insight into mechanisms of stress response, cellular differentiation and disease susceptibility. The principle challenge, however, is that cellular networks remain unknown for most conditions and must be inferred from activity levels of genes (mRNA levels) under different conditions. This dissertation aims to develop computational approaches for inferring, analyzing and validating cellular networks of …

A Neural Network Approach To Border Gateway Protocol Peer Failure Detection And Prediction, Cory B. White

Master's Theses

The size and speed of computer networks continue to expand at a rapid pace, as do the corresponding errors, failures, and faults inherent within such extensive networks. This thesis introduces a novel approach to interface Border Gateway Protocol (BGP) computer networks with neural networks to learn the precursor connectivity patterns that emerge prior to a node failure. Details of the design and construction of a framework that utilizes neural networks to learn and monitor BGP connection states as a means of detecting and predicting BGP peer node failure are presented. Moreover, this framework is used to monitor a BGP network …

Dataset Threshold For The Performance Estimators In Supervised Machine Learning Experiments, Zanifa Omary, Fredrick Mtenzi

Conference papers

The establishment of dataset threshold is one among the first steps when comparing the performance of machine learning algorithms. It involves the use of different datasets with different sample sizes in relation to the number of attributes and the number of instances available in the dataset. Currently, there is no limit which has been set for those who are unfamiliar with machine learning experiments on the categorisation of these datasets, as either small or large, based on the two factors. In this paper we perform experiments in order to establish dataset threshold. The established dataset threshold will help unfamiliar supervised …

Automatic Red Tide Detection Using Modis Satellite Images, Wijian Cheng

USF Tampa Graduate Theses and Dissertations

Red tides pose a significant economic and environmental threat in the Gulf of Mexico. Detecting red tide is important for understanding this phenomenon. In this thesis, machine learning approaches based on Random Forests, Support Vector Machines and K-Nearest Neighbors have been evaluated for red tide detection from MODIS satellite images. Detection results using machine learning algorithms were compared to ship collected ground truth red tide data. This work has three major contributions. First, machine learning approaches outperformed two of the latest thresholding red tide detection algorithms based on bio-optical characterization by more than 10% in terms of F measure and …

A Unified Multi-Functional Dynamic Spectrum Access Framework: Tutorial, Theory And Multi-Ghz Wideband Testbed, Robert Qiu, Nan Guo, Husheng Li, Zhiqiang Wu, Vasu Chakravarthy, Yu Song, Zhen Hu, Peng Zhang, Zhe Chen

Electrical Engineering Faculty Publications

Dynamic spectrum access is a must-have ingredient for future sensors that are ideally cognitive. The goal of this paper is a tutorial treatment of wideband cognitive radio and radar—a convergence of (1) algorithms survey, (2) hardware platforms survey, (3) challenges for multi-function (radar/communications) multi-GHz front end, (4) compressed sensing for multi-GHz waveforms—revolutionary A/D, (5) machine learning for cognitive radio/radar, (6) quickest detection, and (7) overlay/underlay cognitive radio waveforms. One focus of this paper is to address the multi-GHz front end, which is the challenge for the next-generation cognitive sensors. The unifying theme of this paper is to spell out the …

Intentional Learning Agent Architecture, Budhitama Subagdja, Liz Sonenberg, Iyad Rahwan

Research Collection School Of Computing and Information Systems

Dealing with changing situations is a major issue in building agent systems. When the time is limited, knowledge is unreliable, and resources are scarce, the issue becomes more challenging. The BDI (Belief-Desire-Intention) agent architecture provides a model for building agents that addresses that issue. The model can be used to build intentional agents that are able to reason based on explicit mental attitudes, while behaving reactively in changing circumstances. However, despite the reactive and deliberative features, a classical BDI agent is not capable of learning. Plans as recipes that guide the activities of the agent are assumed to be static. …

Predictive Decoding Of Neural Data, Yaroslav O. Halchenko

Dissertations

In the last five decades the number of techniques available for non-invasive functional imaging has increased dramatically. Researchers today can choose from a variety of imaging modalities that include EEG, MEG, PET, SPECT, MRI, and fMRI.

This doctoral dissertation offers a methodology for the reliable analysis of neural data at different levels of investigation. By using statistical learning algorithms the proposed approach allows single-trial analysis of various neural data by decoding them into variables of interest. Unbiased testing of the decoder on new samples of the data provides a generalization assessment of decoding performance reliability. Through consecutive analysis of the …

Application Of Machine Learning Techniques For Real-Time Classification Of Sensor Array Data, Sichu Li

LSU New Orleans Theses and Dissertations

There is a significant need to identify approaches for classifying chemical sensor array data with high success rates that would enhance sensor detection capabilities. The present study attempts to fill this need by investigating six machine learning methods to classify a dataset collected using a chemical sensor array: K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Classification and Regression Trees (CART), Random Forest (RF), Naïve Bayes Classifier (NB), and Principal Component Regression (PCR). A total of 10 predictors that are associated with the response from 10 sensor channels are used to train and test the classifiers. A training dataset of 4 …

Bootstrapping Events And Relations From Text, Ting Liu

Legacy Theses & Dissertations (2009 - 2024)

Information Extraction (IE) is a technique for automatically extracting structured data from text documents. One of the key analytical tasks is extraction of important and relevant information from textual sources. While information is plentiful and readily available, from the Internet, news services, media, etc., extracting the critical nuggets that matter to business or to national security is a cognitively demanding and time consuming task. Intelligence and business analysts spend many hours poring over endless streams of text documents pulling out reference to entities of interest (people, locations, organizations) as well as their relationships as reported in text. Such extracted "information …

Machine Learned Melody Matching Using Strictly Relative Musical Abstractions, Michael Joseph Kolta

Legacy Theses & Dissertations (2009 - 2024)

We implement and evaluate a machine learning approach to improve systems for searching a database of music via melodic sample. We explore symbolic and aural input queries and test our prototypes with extensive user surveys. Our main contribution is to combine the following four elements. First is to create a unique musical abstraction that accounts for both pitch and rhythm in a relative manner. Second, our system allows for approximate matching of imperfect queries via the utilization of the Smith-Waterman algorithm that was originally designed for approximate matching of molecular subsequences, such as DNA samples. Third is to design our …

Quantum Inductive Learning And Quantum Logic Synthesis, Martin Lukac

Dissertations and Theses

Since Quantum Computer is almost realizable on large scale and Quantum Technology is one of the main solutions to the Moore Limit, Quantum Logic Synthesis (QLS) has become a required theory and tool for designing Quantum Logic Circuits. However, despite its growth, there is no any unified aproach to QLS as Quantum Computing is still being discovered and novel applications are being identified.

The intent of this study is to experimentally explore principles of Quantum Logic Synthesis and its applications to Inductive Machine Learning. Based on algorithmic approach, I first design a Genetic Algorithm for Quantum Logic Synthesis that is …

Using Student Mood And Task Performance To Train Classifier Algorithms To Select Effective Coaching Strategies Within Intelligent Tutoring Systems (Its), Robert Sottilare

Electronic Theses and Dissertations

The ultimate goal of this research was to improve student performance by adjusting an Intelligent Tutoring System's (ITS) coaching strategy based on the student's mood. As a step toward this goal, this study evaluated the relationships between each student's mood variables (pleasure, arousal, dominance and mood intensity), the coaching strategy selected by the ITS and the student's performance. Outcomes included methods to increase the perception of the intelligent tutor to allow it to adapt coaching strategies (methods of instruction) to the student's affective needs to mitigate barriers to performance (e.g. negative affect) during the one-to-one tutoring process. The study evaluated …

Concept Learning By Example Decomposition, Sameer Joshi

Electronic Theses and Dissertations

For efficient understanding and prediction in natural systems, even in artificially closed ones, we usually need to consider a number of factors that may combine in simple or complex ways. Additionally, many modern scientific disciplines face increasingly large datasets from which to extract knowledge (for example, genomics). Thus to learn all but the most trivial regularities in the natural world, we rely on different ways of simplifying the learning problem. One simplifying technique that is highly pervasive in nature is to break down a large learning problem into smaller ones; to learn the smaller, more manageable problems; and then to …

Analyses Of Crash Occurence And Injury Severities On Multi Lane Highways Using Machine Learning Algorithms, Abhishek Das

Electronic Theses and Dissertations

Reduction of crash occurrence on the various roadway locations (mid-block segments; signalized intersections; un-signalized intersections) and the mitigation of injury severity in the event of a crash are the major concerns of transportation safety engineers. Multi lane arterial roadways (excluding freeways and expressways) account for forty-three percent of fatal crashes in the state of Florida. Significant contributing causes fall under the broad categories of aggressive driver behavior; adverse weather and environmental conditions; and roadway geometric and traffic factors. The objective of this research was the implementation of innovative, state-of-the-art analytical methods to identify the contributing factors for crashes and injury …

Real-Time Automatic Price Prediction For Ebay Online Trading, Ilya Igorevitch Raykhel

Theses and Dissertations

While Machine Learning is one of the most popular research areas in Computer Science, there are still only a few deployed applications intended for use by the general public. We have developed an exemplary application that can be directly applied to eBay trading. Our system predicts how much an item would sell for on eBay based on that item's attributes. We ran our experiments on the eBay laptop category, with prior trades used as training data. The system implements a feature-weighted k-Nearest Neighbor algorithm, using genetic algorithms to determine feature weights. Our results demonstrate an average prediction error of 16%; …

Development Of A Workflow For The Comparison Of Classification Techniques, Zanifa Omary

Masters

As the interest in machine learning and data mining springs up, the problem of how to assess learning algorithms and compare classifiers become more pressing. This has been associated with the lack of comprehensive and complete workflow depending on the project scale to provide guidance to its users. This means the success or failure of the project can be highly dependent on the person or team carrying it. The standard practice adopted by many researchers and experimenters has been to follow steps or phases from existing workflows such as CRISP-DM, KDD and SASSEMMA. However, as machine learning and data mining …

Clustering Via Supervised Support Vector Machines, Sepehr Merat

LSU New Orleans Theses and Dissertations

An SVM-based clustering algorithm is introduced that clusters data with no a priori knowledge of input classes. The algorithm initializes by first running a binary SVM classifier against a data set with each vector in the set randomly labeled. Once this initialization step is complete, the SVM confidence parameters for classification on each of the training instances can be accessed. The lowest confidence data (e.g., the worst of the mislabeled data) then has its labels switched to the other class label. The SVM is then re-run on the data set (with partly re-labeled data). The repetition of the above process …

Classification Algorithms On The Cell Processor, Mateusz Wyganowski

Theses

The rapid advancement in the capacity and reliability of data storage technology has allowed for the retention of virtually limitless quantity and detail of digital information. Massive information databases are becoming more and more widespread among governmental, educational, scientific, and commercial organizations. By segregating this data into carefully defined input (e.g.: images) and output (e.g.: classification labels) sets, a classification algorithm can be used develop an internal expert model of the data by employing a specialized training algorithm. A properly trained classifier is capable of predicting the output for future input data from the same input domain that it was …

Scaling Ant Colony Optimization With Hierarchical Reinforcement Learning Partitioning, Erik J. Dries, Gilbert L. Peterson

Faculty Publications

This paper merges hierarchical reinforcement learning (HRL) with ant colony optimization (ACO) to produce a HRL ACO algorithm capable of generating solutions for large domains. This paper describes two specific implementations of the new algorithm: the first a modification to Dietterich’s MAXQ-Q HRL algorithm, the second a hierarchical ant colony system algorithm. These implementations generate faster results, with little to no significant change in the quality of solutions for the tested problem domains. The application of ACO to the MAXQ-Q algorithm replaces the reinforcement learning, Q-learning, with the modified ant colony optimization method, Ant-Q. This algorithm, MAXQ-AntQ, converges to solutions …

Detection Of Marine Vehicles In Images And Video Of Open Sea, Sergiy Fefilatyev

USF Tampa Graduate Theses and Dissertations

This work presents a new technique for automatic detection of marine vehicles in images and video of open sea. Users of such system include border guards, military, port safety, flow management, and sanctuary protection personnel. The source of images and video is a digital camera or a camcorder which is placed on a buoy or stationary mounted in a harbor facility. The system is intended to work autonomously, taking images of the surrounding ocean surface and analyzing them for the presence of marine vehicles. The goal of the system is to detect an approximate window around the ship. The proposed …

Assessing The Costs Of Sampling Methods In Active Learning For Annotation, James Carroll, Robbie Haertel, Peter Mcclanahan, Eric K. Ringger, Kevin Seppi

Faculty Publications

Traditional Active Learning (AL) techniques assume that the annotation of each datum costs the same. This is not the case when annotating sequences; some sequences will take longer than others. We show that the AL technique which performs best depends on how cost is measured. Applying an hourly cost model based on the results of an annotation user study, we approximate the amount of time necessary to annotate a given sentence. This model allows us to evaluate the effectiveness of AL sampling methods in terms of time spent in annotation. We acheive a 77% reduction in hours from a random …

Comparison Of Machine Learning Algorithms For Modeling Species Distributions: Application To Stream Invertebrates From Western Usa Reference Sites, Margi Dubal

All Graduate Plan B and other Reports, Spring 1920 to Spring 2023

Machine learning algorithms are increasingly being used by ecologists to model and predict the distributions of individual species and entire assemblages of sites. Accurate prediction of distribution of species is an important factor in any modeling. We compared prediction accuracy of four machine learning algorithms-random forests, classification trees, support vector machines, and gradient boosting machines to a traditional method, linear discriminant models (LDM), on a large set of stream invertebrate data collected at 728 reference sites in the western United States. Classifications were constructed for individual species and for assemblages of sites clustered a priori by similarity on biological characteristics. …

Improving Liquid State Machines Through Iterative Refinement Of The Reservoir, R David Norton

Theses and Dissertations

Liquid State Machines (LSMs) exploit the power of recurrent spiking neural networks (SNNs) without training the SNN. Instead, a reservoir, or liquid, is randomly created which acts as a filter for a readout function. We develop three methods for iteratively refining a randomly generated liquid to create a more effective one. First, we apply Hebbian learning to LSMs by building the liquid with spike-time dependant plasticity (STDP) synapses. Second, we create an eligibility based reinforcement learning algorithm for synaptic development. Third, we apply principles of Hebbian learning and reinforcement learning to create a new algorithm called separation driven synaptic modification …