Theses and Dissertations at Montana State University (MSU)
Permanent URI for this collectionhttps://scholarworks.montana.edu/handle/1/733
Browse
9 results
Search Results
Item Developing bio-inspired methodologies for encoding angular position from strain(Montana State University - Bozeman, College of Engineering, 2020) Lange, Christopher William; Chairperson, Graduate Committee: Mark JankauskiAs mechanical systems rely more on closed-loop control, the sensors which supply feedback information are essential. Additionally, in systems where sensor function is critical, sensor redundancy is important to retain functionality if one or more sensors fail. Redundancy can be achieved through multiple high-fidelity sensors which measure the same type of information, such as gyroscopes or accelerometers. However, multiple high-fidelity sensors can increase cost significantly. This thesis explores the potential to replace or augment the functionality of angular position sensors using strain measurements. Strain gauges are already used in system health monitoring systems. By utilizing these already implemented sensors to measure angular position, we can remove the additional cost of redundant angular position sensors. However, for complex systems, the mapping between strain and angular position is unclear. By incorporating reduced order, physics-based models into machine learning techniques, we can efficiently transform high-order strain data into angular position. To demonstrate the potential of using alternative sensing methods, we developed a reduced order model of a parametrically excited flexible pendulum. Inspiration for this simplified system comes from insect halteres, which are small sensory organs evolved from insect hind wings which provide rapid information about body rotation. The parametrically excited flexible pendulum allows a single axis of rotation and single direction of flexibility to be paired, and their relationship studied. By varying parameters within the model such as pendulum length and modulus as well as parametric excitation amplitude and frequency, the Gaussian process regression learning can be optimized to reduce training time and increase untrained prediction accuracy. Inputs of strain and parametric excitation position along with their respective first and second derivatives are then analyzed to determine which inputs are interrelated and therefore un-necessary, thus reducing the input required. This provides the essential first steps towards using machine learning to implement multiple sensor, deformation based, multi axial angular position sensing in complex systems.Item Mapping risky driver behavior and identifying their contributory factors : a spatial statistical approach(Montana State University - Bozeman, College of Engineering, 2016) Sharda, Shivam; Chairperson, Graduate Committee: Yiyi WangThe goal of this study is to develop risk maps that predict zones that have an increased risk of traffic crashes for utility service trucks of the electric power industry. This study employed a national dataset that contains video logs of driving events from 10,009 utility truck drivers from 2010 to 2014. The study consists of four steps. Step 1 focused on finding whether certain driver behaviors (e.g., traffic violation, distraction, etc.) cluster in the same location as crashes and, therefore, suggest behaviors that are predictive of crashes. The study used Getis-Ord Gi* hot-spot analysis to reveal the clustering pattern within a standard unit of area (at the grid cell level: 1,640 feet by 1,640 feet). The finding of this research indicated that four behaviors ('risky behaviors') consistently cluster with the collision outcomes: distraction, lack of awareness, following too close, and eating/drinking. In Step 2, negative binomial models were used to relate the occurrence of the risky behaviors to a host of geospatial variables (e.g., land use, traffic, and socio-economics) while controlling for the exposure at the grid level (200 feet by 200 feet, roughly the size of a street block). Step 2 was implemented on the three datasets that were assembled based on different levels of availability of the geospatial features. Results indicated that well-balanced land use, road network density, lane-mile density of secondary and primary roads (if urban areas), and high concentration of elderly people (65 years and above) contributed to the prevalence of risky behaviors. Residential neighborhood, local road (if rural area), and average household size were shown to dampen incidence of risky behaviors. Step 3 developed the scoring systems to estimate the overall risk of each risky behavior for a given location (grid). Finally, Step 4 developed risk maps on a 2-D scale to delineate locations into different levels of hazards. In sum, this study confirmed the linkage between driver behavior and collisions and proposed a new way to anticipate crashes. While the test dataset pertains to utility service trucks, the methods can be adapted for predicting locations where the risk of future crashes is higher.Item Is model averaging the solution for addressing model uncertainty? : methodological insights, tools for assessment, and considerations for practical use(Montana State University - Bozeman, College of Letters & Science, 2016) Banner, Katharine Michelle; Chairperson, Graduate Committee: Steve Cherry; Megan Higgs (co-chair); Megan Higgs was a co-author of the article, 'Considerations for assessing model averaging of regression coefficients' in the journal 'Ecological applications' which is contained within this thesis.; Megan Higgs was a co-author of the article, 'Investigating the posterior variance of partial regression coefficients resulting from three common methods for multimodel inference' submitted to the journal 'Annals of applied statistics' which is contained within this thesis.; Megan Higgs was a co-author of the article, 'The model averaged posteriors plot MAPP package for R statistical software' submitted to the journal 'Journal of statistical software' which is contained within this thesis.Model averaging (MA) was developed as a way to combine predictions from many models, with the goal of reducing bias and incorporating model uncertainty into final predictive inferences. A new flavor of MA, focused on averaging partial regression coefficients over multiple models, has gained traction in fields such as Ecology, Biology, and Political Science, with motivation stemming from the concern that inferences based on a single model are too 'naive' (i.e., do not fairly reflect sources of substantial uncertainty). However, coefficients appearing in multiple models do not necessarily hold the same interpretation across models, and averaging over them has the potential to result in inferences that are difficult to interpret. A gap exists between the theoretical development of MA and its current use in practice, potentially leaving well-intentioned researchers with unclear inferences, or difficulties justifying decisions to use (or not use) MA. Furthermore, it is questionable whether the perceived benefit of accounting for an additional source of uncertainty is realized in terms of increased variance for quantities of interest. In this work, we revisit relevant foundations of regression modeling, suggest more explicit notation and graphical tools, and discuss how individual model results are combined to obtain a MA result, with the goal of helping researchers make informed decisions about MA. We present a new package for R Statistical Software providing plotting functions for visualizing components going into the MA posterior distribution. This package is meant to be used to assess the implicit assumptions made by using MA for regression coefficients, complete with guidelines for use and examples. We also design and conduct a simulation study to investigate how the variance for a partial regression coefficient of interest is different for three different approaches used within multimodel inference (MA using all models, MA using a subset of models, and conditioning inferences on one model). We assess whether the perceived benefit of accounting for model uncertainty is actually realized when more models are used for final inference, with the goal of helping researches weigh tradeoffs between using variants of MA in place of one well thought out model.Item Subsidizing strippers : the impact of royalty rate reductions on the intensive and extensive production margins of marginally producing oil wells(Montana State University - Bozeman, College of Agriculture, 2016) Bishop, Zachary Andrew; Chairperson, Graduate Committee: Randal R. RuckerSubstantial research has been conducted on the impacts of taxation on oil production. However, a void in the literature exists as the distinction has not yet been made between significant and marginal oil production. Using well-level production data from the state of Wyoming, this thesis estimates the impact of royalty rate reductions on marginally-producing, federal oil wells--commonly referred to as stripper wells. The empirical analysis is conducted using fixed effects double- and triple-difference models and a more traditional multiple regression model. The results suggest that production from federal stripper wells increased substantially during the royalty rate reduction program--on both the intensive and extensive production margins.Item Anderson-Darling Regression with two examples from biofilm engineering(Montana State University - Bozeman, College of Letters & Science, 1997) Daly, Don SimoneItem Predictive models for 30-day patient readmissions in a small community hospital(Montana State University - Bozeman, College of Engineering, 2013) Lovejoy, Matthew Walter; Chairperson, Graduate Committee: David ClaudioPresently, national healthcare initiatives have a strong emphasis on improving patient quality of care through a reduction in patient readmissions. Current federal regulations created through the Patient Protection and Affordable Care Act (PPACA); focus on the reduction in readmissions to improve patient quality of care (Stone & Hoffman, 2010). This legislation mandates decreased reimbursement for services if a facility has high 30-day patient readmissions related to the core measures Congestive Heart Failure (CHF), Acute Myocardial Infarction (AMI) and Pneumonia (PNM). This research focuses on building predictive models to aid Bozeman Deaconess Health Services (BDHS), a small community hospital, reduce their readmission rates. Assistance was performed through identification of patient characteristics influencing patient readmission risk, along with advanced statistical regression techniques used to develop readmission risk prediction models. Potential predictor variables and prediction models were obtained through retrospective analysis of patient readmission data from BDHS during January 2009 through December 2010. For increased prediction accuracy seven separate readmission dataset types were developed: General population, and ICD-9 code related populations for AMI, CHF, PNM, Alcoholic Psychoses (291), Cardiac Dysrhythmias (427) and Disorders of the Function of the Stomach (536). For the greatest benefit from readmission reduction, analysis focused on readmissions categorized as Potentially Preventable Readmissions (PPR); defined as unplanned, medically related readmissions within 30-days of a patient's previous inpatient visit. General exploratory analysis was performed on the PPR patient data to discover patterns which may indicate certain variables as good predictors of patient readmission risk. The prediction model methods compared were binary logistic regression, and multivariate adaptive regression splines (MARS). Usable binary logistic regression models for 536 (Nagelkerke R 2=0.676) and CHF (Nagelkerke R 2=0.974) were achieved. MARS developed usable models for 427 (Naïve Adj. R 2=0.63288), 536 (Naïve Adj. R 2=0.77395), AMI (Naïve Adj. R 2=0.76705), CHF (Naïve Adj. R 2=0.99385) and PNM (Naïve Adj. R 2=0.82615). Comparison of the modeling methods suggest MARS is more accurate at developing usable prediction models, however a tradeoff between model complexity and predictability is present. The usable readmission risk prediction models developed for BDHS will aid BDHS in reducing their readmissions rates, consequently improving patient quality of care.Item The effect of basin physiography on the spatial distribution of snow water equivalent and snow density near peak accumulation by Karl Bruno Wetlaufer.(Montana State University - Bozeman, College of Letters & Science, 2013) Wetlaufer, Karl Bruno; Chairperson, Graduate Committee: Jordy HendrikxThis study quantifies the effect of the physiography (elevation, potential incoming solar radiation, land cover, etc.) of a large (207 km 2) and complex mountainous basin on the spatial distribution of snow water equivalence (SWE) and snow density during peak SWE accumulation. SWE and snow density were sampled in areas of the basin that were physiographically representative (based on unique combinations of elevation, incoming solar radiation, and land cover) to the basin as a whole. Sampling took place over a variety of spatial scales (10m-400m) in a semi-random and structured manner acquiring over 1,000 direct measurements of SWE and snow density. Three modeling approaches were used in the analysis of the SWE data; regression tree, conditional inference tree, and mixed effects multiple regression. The three modeling approaches were similar in their estimates of total basin SWE (approximately within 1% of their averages) but provided very different patterns of how SWE is spatially distributed throughout the basin. All three methods showed elevation and potential incoming solar radiation to have the most significant influence on the spatial distribution of SWE, with land cover also being significant in the mixed effects and conditional inference tree models. Snow density was observed to vary widely throughout the basin with a standard deviation of 61 kg/m 3 around a mean of 349 kg/m 3. The spatial distribution of density was modeled using regression tree and multiple linear regression analysis. Both models estimated similar basin average snow density using elevation and radiation as explanatory variables, but displayed considerably different spatial distributions and ranges of value. This study demonstrates the importance of elevation and radiation for modeling the spatial distribution of SWE and snow density in a large and physiographically diverse basin and expresses the differences that exist between various methods of modeling these phenomena.Item An economic analysis of the impact of decoupled payments on farm solvency in the United States(Montana State University - Bozeman, College of Agriculture, 2014) Hasenoehrl, Amy Rae; Co-chairs, Graduate Committee: Eric Belasco and Anton BekkermanThis thesis evaluates the effects of decoupled agricultural support payments on the debt-to-asset ratio of farmers in the top five states producing corn, cotton, wheat and soybeans from 1996 to 2011. Building on existing literature, this study estimates the broader impacts of decoupled payments on farm solvency by considering all decoupled payments made since their establishment in 1996. A theoretical model of profit maximization identifies the factors predicted to influence solvency, which include farm assets, income, expenses, scale, production risk, decoupled payments and operator characteristics. Following the literature, the relationship between these factors and farm solvency are estimated empirically using a linear regression model with data from the Agricultural Resource Management Survey, Farm Service Agency and Risk Management Agency. The results indicate decoupled payments have a positive relationship with the debt-to-asset ratio and that the elimination of decoupled payments in the upcoming Farm Bill could lead to decreases in farmers' debt-to-asset ratios by an average of approximately ten percent. Furthermore, an analysis of the effects of decoupled payments by primary crop designation suggests that only corn soybean, corn and wheat farmers' debt-to-asset ratios are significantly responsive to changes in decoupled payments. This study also finds the effect of decoupled payments on solvency is uniform across farm size. In addition to these results, this thesis also contributes to current literature by providing preliminary evidence of an endogenous relationship between acres operated and the debt-to-asset ratio, which appears to introduce a positive bias on the parameter estimate for decoupled payments in the linear regression model. Furthermore, when a two-stage least squares model is used to control for this bias, the results estimate a negative relationship between decoupled payments and the debt-to-asset ratio. Due to the change in the coefficient of decoupled payments between the two models, this study suggests that results from research failing to account for a potential endogenous relationship between acres and the debt-to-asset ratio should be interpreted with caution.Item Empirical least squares regression models for employment, in- and out-migration, and income distribution in the Northern Great Plains region of the United States(Montana State University - Bozeman, College of Agriculture, 1974) Lewis, Eugene Patrick, 1948-; Chairperson, Graduate Committee: Lloyd D. Bender.This research effort is aimed at determining empirical least squares regression models for employment, in- and out-migration, and income distribution, Secondary data is used exclusively; The observations are 181 non-metro counties in the Northern Great Plains Region of the United States. The statistical results show that all four models are directly linked to variations in the economic bases of counties. To some extent, this allows the models to be used concurrently in determining impacts. It was hypothesized and shown that the multiplier effect for employment varies with industry, scale of operation of the various industries, and location in economic space. This conslusion along with the successful inclusion of migration and income distribution suggests that the approach taken in this study is a possible alternative to strandard aggregate economic base and input-output studies.