Statistical Models Commons^™

Analyzing Relationships With Machine Learning, Oscar Ko

Dissertations, Theses, and Capstone Projects

Procedurally, this project aims to take a dataset, analyze it, and offer insights to the audience in an easy-to-digest format. Conceptually, this project will seek to explore questions like: “Do couples that meet through online dating or dating apps have higher or lower quality relationships?”, “Can any features in this dataset help predict how a subject would rate their relationship quality?”, and “What other insights can I derive from using machine learning for exploratory analysis?” The intended audience for this project is anyone interested in romantic relationships or machine learning.

The dataset is from a Stanford University survey, “How Couples …

Biasing Estimator To Mitigate Multicollinearity In Linear Regression Model, Abdulrasheed Bello Badawaire, Issam Dawoud, Adewale Folaranmi Lukman, Victoria Laoye, Arowolo Olatunji

Al-Bahir

A new two-parameter estimator was developed to combat the threat of multicollinearity for the linear regression model. Some necessary and sufficient conditions for the dominance of the proposed estimator over ordinary least squares (OLS) estimator, ridge regression estimator, Liu estimator, KL estimator, and some two-parameter estimators are obtained in the matrix mean square error sense. Theory and simulation results show that, under some conditions, the proposed two-parameter estimator consistently dominates other estimators considered in this study. The real-life application result follows suit.

Statistical Tolerance Regions For Flexible Modeling Paradigms, Yafan Guo

Theses and Dissertations--Statistics

Tolerance intervals in a regression setting allow the user to quantify, with a specified degree of confidence, bounds for a specified proportion of the sampled population when conditioned on a set of covariate values. While methods are available for tolerance intervals in fully-parametric regression settings, the construction of tolerance intervals for semiparametric regression models has been treated in a limited capacity. The first project fills this gap and develops likelihood-based approaches for the construction of pointwise one-sided and two-sided tolerance intervals for semiparametric regression models. A numerical approach is also presented for constructing simultaneous tolerance intervals. An appealing facet of …

Forecasting Remission Time Of A Treatment Method For Leukemia As An Application To Statistical Inference Approach, Ahmed Galal Atia, Mahmoud Mansour, Rashad Mohamed El-Sagheer, B. S. El-Desouky

Basic Science Engineering

In this paper, Weibull-Linear Exponential distribution (WLED) has been investigated whether being it is a well-fit distribution to a clinical real data. These data represent the duration of remission achieved by a certain drug used in the treatment of leukemia for a group of patients. The statistical inference approach is used to estimate the parameters of the WLED through the set of the fitted data. The estimated parameters are utilized to evaluate the survival and hazard functions and hence assessing the treatment method through forecasting the duration of remission times of patients. A two-sample prediction approach has been applied to …

Classification Of Adult Income Using Decision Tree, Roland Fiagbe

Data Science and Data Mining

Decision tree is a commonly used data mining methodology for performing classification tasks. It is a tree-based supervised machine learning algorithm that is used to classify or make predictions in a path of how previous questions are answered. Generally, the decision tree algorithm categorizes data into branch-like segments that develop into a tree that contains a root, nodes, and leaves. This project seeks to explore the decision tree methodology and apply it to the Adult Income dataset from the UCI Machine Learning Repository, to determine whether a person makes over 50K per year and determine the necessary factors that improve …

Aircraft Damage Classification By Using Machine Learning Methods, Tüzün Tolga İnan

International Journal of Aviation, Aeronautics, and Aerospace

Safety is the most significant factor that affected incidents (non-fatal) and accidents (fatal) in civil aviation history related to scheduled flights. In the history of scheduled flights, the total incident and accident number until 2022 is 1988. In this study, 677 of them are taken into consideration since 11 September 2001. The purpose of this study is to reveal the factors that can classify type of aircraft damages such as none, minor and substantial in all-time incidents and accidents. ML algorithms with different configurations are applied for the classification process. The RFE and PCA are used to find the most …

Stochastic Optimization To Reduce Aircraft Taxi-In Time At Igia, New Delhi, Rajib Das, Saileswar Ghosh, Rajendra Desai, Pijus Kanti Bhuin, Stuti Agarwal

International Journal of Aviation, Aeronautics, and Aerospace

Since there is an uncertainty in the arrival times of flights, pre-scheduled allocation of runways and stands and the subsequent first-come-first-served treatment results in a sub-optimal allocation of runways and stands, this is the prime reason for the unusual delays in taxi-in times at IGIA, New Delhi.

We simulated the arrival pattern of aircraft and utilized stochastic optimization to arrive at the best runway-stands allocation for a day. Optimization is done using a GRG Non-Linear algorithm in the Frontline Systems Analytic Solver platform. We applied this model to eight representative scenarios of two different days. Our results show that without …

The Influence Of Urban Forms And Street Infrastructure On Pedestrian-Motorist Collisions, Taylor J. Foreman

College of Graduate Studies: Theses & Dissertations

Unwalkable cities are afflicted by serious issues such as increasing rates of pedestrian traffic accidents, public health concerns, and the denied right to have an accessible city. This study examines how different types of urban forms and street infrastructure contribute to the prevalence of traffic accidents in two major metropolitan cities in the United States: Atlanta, Georgia, and Boston, Massachusetts. This study utilizes geospatial analysis through the Average Nearest Neighbor and Optimized Hot Spot Analysis tools to determine the spatial distribution of traffic accidents throughout both cities. Additionally, statistical tests were conducted to explore the relationships between the number of …

Bayesian Structural Time Series Methods For Modeling Cattle Body Temperature In Heat-Stressed Animals, Lacey Quandt

Murray State Theses and Dissertations

Climate change has had devastating effects globally, most commonly talked about during natural disasters and rising temperatures. Notably, the climate concern is turning towards agriculture and livestock. With rising temperatures, the prolonged amount of heat stress put on animals, specifically cattle, is becoming more apparent. Heat stress has been linked to a reduction in cattle growing and fattening, feed intake, productivity, reproduction, and fertility; increased heart rates and respiration; changes in behavior; and mortality in severe cases. There are abatement strategies put in place to lower heat stress in cattle, such as improvements in shading and cooling, nutritional management, and …

Statistical Intervals For Neural Network And Its Relationship With Generalized Linear Model, Sheng Yuan

Theses and Dissertations--Statistics

Neural networks have experienced widespread adoption and have become integral in cutting-edge domains like computer vision, natural language processing, and various contemporary fields. However, addressing the statistical aspects of neural networks has been a persistent challenge, with limited satisfactory results. In my research, I focused on exploring statistical intervals applied to neural networks, specifically confidence intervals and tolerance intervals. I employed variance estimation methods, such as direct estimation and resampling, to assess neural networks and their performance under outlier scenarios. Remarkably, when outliers were present, the resampling method with infinitesimal jackknife estimation yielded confidence intervals that closely aligned with nominal …

High Dimensional Data Analysis: Variable Screening And Inference, Lei Fang

Theses and Dissertations--Statistics

This dissertation focuses on the problem of high dimensional data analysis, which arises in many fields including genomics, finance, and social sciences. In such settings, the number of features or variables is much larger than the number of observations, posing significant challenges to traditional statistical methods.

To address these challenges, this dissertation proposes novel methods for variable screening and inference. The first part of the dissertation focuses on variable screening, which aims to identify a subset of important variables that are strongly associated with the response variable. Specifically, we propose a robust nonparametric screening method to effectively select the predictors …

Novel Modelling And Inference Considerations Involving The Exponentially-Modified Gaussian Distribution, Yanxi Li

Theses and Dissertations--Statistics

The exponentially-modified Gaussian (EMG) distribution is well-suited for analyzing data with positive skewness due to its characteristic positive skew from the exponential component. Despite its popularity in various fields, the EMG distribution has only been analyzed for univariate data without any regression settings. To address this limitation, we developed a generalized EMG regression model with covariates by assigning parametric functional forms to some or all of the parameters in the EMG distribution that vary with values of the covariates. To further perform data-clustering on observation points, we propose a competing regression model where the error structure is assumed to be …

Finite Mixtures Of Mean-Parameterized Conway-Maxwell-Poisson Models, Dongying Zhan

Theses and Dissertations--Statistics

For modeling count data, the Conway-Maxwell-Poisson (CMP) distribution is a popular generalization of the Poisson distribution due to its ability to characterize data over- or under-dispersion. While the classic parameterization of the CMP has been well-studied, its main drawback is that it is does not directly model the mean of the counts. This is mitigated by using a mean-parameterized version of the CMP distribution. In this work, we are concerned with the setting where count data may be comprised of subpopulations, each possibly having varying degrees of data dispersion. Thus, we propose a finite mixture of mean-parameterized CMP distributions. An …

Methodologies And Computational Tools For Zero-Inflated Discrete Weibull Models, Peng Yeh

Theses and Dissertations--Statistics

Count data with excess zeros is common in many fields, such as ecology, healthcare, and insurance. Excess zeros data are often causing the inaccurate fit from the count models. While zero-inflated models have been developing for over two decades, one should also consider a more flexible model that can handle the excess zeros and further over- or under-dispersion. In this talk, we discuss zero-inflated discrete Weibull model and some novel computational contributions. The flexibility and competitiveness of the ZIDW model are illustrated by simulation studies and a real data analysis. We also investigate the performance of the proposed model through …

Potential Alzheimer's Disease Plasma Biomarkers, Taylor Estepp

Theses and Dissertations--Epidemiology and Biostatistics

In this series of studies, we examined the potential of a variety of blood-based plasma biomarkers for the identification of Alzheimer's disease (AD) progression and cognitive decline. With the end goal of studying these biomarkers via mixture modeling, we began with a literature review of the methodology. An examination of the biomarkers with demographics and other health factors found evidence of minimal risk of confounding along the causal pathway from biomarkers to cognitive performance. Further study examined the usefulness of linear combinations of biomarkers, achieved via partial least squares (PLS) analysis, as predictors of various cognitive assessment scores and clinical …

The Impact Of Subjective Risk Analysis On Real Estate Prices In The Nisqually Region Following The 2001 Nisqually Earthquake, Ryan Espedal

All Master's Theses

Earthquakes are an environmental hazard that pose great risks to communities almost every day. With earthquakes, the main cause of concern is physical destruction of property, however, there are also psychological effects that are researched and discussed much less. In 2001, the Nisqually area of western Washington experienced a substantial earthquake that produced minimal physical damage but caused a significant decrease in real estate prices. Studying single-family homes from 1986-2012, this research utilizes hedonic property models to measure the change in consumer’s subjective risk calculations with reference to real estate purchases after the Nisqually earthquake, measure the relationship between earthquake …

Carnivore And Ungulate Occurrence In A Fire-Prone Region, Sara J. Moriarty-Graves

Cal Poly Humboldt theses and projects

Increasing fire size and severity in the western United States causes changes to ecosystems, species’ habitat use, and interspecific interactions. Wide-ranging carnivore and ungulate mammalian species and their interactions may be influenced by an increase in fire activity in northern California. Depending on the fire characteristics, ungulates may benefit from burned habitat due to an increase in forage availability, while carnivore species may be differentially impacted, but ultimately driven by bottom-up processes from a shift in prey availability. I used a three-step approach to estimate the single-species occupancy of four large mammal species: mountain lion (Puma concolor), coyote …

An Assessment Of "Long-Thin" Airline Routes: Network Structure And Emissions Implications For Environmental Policy, Porter Burns

All Master's Theses

The purpose of this research was to define, map, and quantify the network and environmental implications of “long-thin” routes (LTRs) – a route structure that has been discussed in the aviation industry but not formally studied in literature. LTRs were defined through the use of global OAG scheduling data from 1998 to 2018 to identify trends in air traffic growth and network dynamics. Flights were separated into seven aircraft class sizes (e.g., 75–150 seats, 150–225 seats) to measure LTRs at multiple scales. Routes were considered “long” if the stage length was at or above the 75^th percentile in each …

Novel Feature Evaluation In Ultra-High Dimensional Right-Censored Data, With Applications To Head And Neck Cancer, Atika Farzana Urmi

Graduate Research Posters

Background: Head and neck cancer is the 6th most common cancer worldwide with an expected 1.08 million new cases each year. Such cancer data are ultra-high dimensional with thousands of clinical features and gene expressions, making it challenging for the traditional analytical tools to extract the potential biomarker for the cancer survival and control false discoveries. In addition, presence of heavy censoring can affect the screening procedures based on Kaplan-Meier (K-M) survival estimates.

Aim: To propose a model free, ultra-high dimensional feature screening method with two-dimensional survival outcome allowing false discovery rate (FDR) control.

Method: 516 primary tumor patients with …

Applications Of Transfer Learning From Malicious To Vulnerable Binaries, Sean Patrick Mcnulty

Graduate Student Theses, Dissertations, & Professional Papers

Malware detection and vulnerability detection are important cybersecurity tasks. Previous research has successfully applied a variety of machine learning methods to both. However, despite their potential synergies, previous research has yet to unite these two tasks. Given the recent success of transfer learning in many domains, such as language modeling and image recognition, this thesis investigated the use of transfer learning to improve vulnerability detection. Specifically, we pre-trained a series of models to detect malicious binaries and used the weights from those models to kickstart the detection of vulnerable binaries. In our study, we also investigated five different data representations …