Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 18 of 18 for “"Boosted Regression Trees"”.
-
Data-driven prediction of saltmarsh morphodynamics
… to which machine learning model approaches (boosted regression trees, neural networks and Bayesian networks) can facilitate synthesis of information and prediction of decadal-scale morphological tendencies of saltmarshes. Importantly, data-driven predictions are independent of the assumptions …
-
Regional data refine local abundance models: modeling plant species abundance distributions on the Central Plains
… data to regional scale landscapes. Using boosted regression trees, I examined the issues of spatial scale and errors associated with extrapolating species distribution models developed using locally collected abundance data to regional extents for a native and alien plant species across a …
-
Can ecological models be transferred in time and space? An investigation using a range of contrasting taxa
… three popular correlative modelling algorithms; boosted regression trees (BRTs), generalised linear models (GLMs) and Maxent, before I investigate the transferability of a further five algorithms and an ensemble approach. Furthermore, I investigate the transferability of a simple and potentially …
-
Assessing predictive performance and transferability of species distribution models for freshwater fish in the United States
… validation tended to be higher in the logistic regression and boosted regression trees (BRT) than the presence-only MaxEnt models. AUC in the cross validation was also higher for species with large geographic ranges and small local populations. Species prevalence affected discrimination power in …
-
Lyme Disease and Forest Fragmentation in the Peridomestic Environment
… models of Lyme in the NRHD. We did so using both Boosted Regression Trees (BRT) and Maximum Entropy (MaxEnt) modeling, the two most common niche modeling algorithms in the field today. We found that Lyme is strongly associated with higher density of developed-herbaceous edges within 100-meters …
-
Streams in a changing landscape : identifying candidate reference reaches to assess the physical and biotic integrity of Missouri's wadeable streams
… area, reach gradient, spring density), we used boosted regression trees to model the influence of channel morphology, substrate, cover, and water quality, and watershed-level flow modification and fragmentation, urbanization, agriculture, and point-source pollution on stream fish and aquatic …
-
Flow and thermal regimes in river networks: effects of hydropower regulation and climate extremes
… Based on the machine learning technique of boosted regression trees (BRT), spatial distribution of water residence time is estimated for the long-term annual average hydrological conditions and extreme cases of flood and drought. • Impacts of hydropower over temporal and spatial range are …
-
Remote sensing-based analysis of land cover detrimental for ecosystem services in fragile lands of Ethiopia
… approach. Undercover cropland was mapped using Boosted Regression Trees on field observed percent cover, topographic and location and parameters. The influential factors of undercover cropland are elevation, distance to settlements, slope, East aspect and distance to national park with elevation …
-
Examining Atmospheric and Ecological Drivers of Wildfires, Modeling Wildfire Occurrence in the Southwest United States, and Using Atmospheric Sounding Observations to Verify National Weather Service Spot Forecasts
… with increased wildfire activity. We implemented boosted regression trees (BRT) to model wildfire occurrence for all and large wildfires for different wildfire types (i.e., lightning, human) across the SWA by PSA. BRT models for all wildfires demonstrated relatively small mean and mean absolute …
-
Exotic Species Invasion and Biodiversity in Bangladesh Forest Ecosystems
… and exotic plant invasion in this ecosystem. Boosted Regression Trees and Detrended Correspondence Analysis are used to determine these relationships. Most exotics are trees followed by shrubs and herbs. Fabaceae and Asteraceae contribute a large proportion of exotic species. Most of them …
-
Red-headed Woodpecker Habitat Selection and Breeding Ecology on Fort Drum Military Installation, New York
… to assess nest-habitat selection.;I used boosted regression trees (BRT) to identify red-headed woodpecker nest-habitat thresholds at multiple spatial scales. Red-headed woodpeckers exhibited nest-tree thresholds related to tree (1) decay class, (2) dead-limb length, and (3) diameter at …
-
Drivers of landscape evolution in northern ecosystems: Shrub expansion, vegetation-fire interaction, and permafrost degradation
… a convolutional neural network and use gradient-boosted regression trees to analyze the environmental drivers of the degradation over the last 70 years. My analysis finds permafrost degradation to increase over time, but that much of this degradation is ephemeral, with older features draining as …
-
Applying new aquatic habitat data to manage invasive and imperiled crayfishes in North America
North America is home to the majority of the world’s crayfish diversity, but many of these crayfishes are considered invasive, and an even greater number are highly imperiled and at risk for extinction. In addition, new GIS data layers developed specifically for freshwater have recently been …
-
Investigating ecological drivers and impacts of vegetation change in sub Antarctic tundra
… of key soil 4 properties across MI using boosted regression trees. The impact of temperature change on plant species distribution was then examined through species distribution models incorporating edaphic properties along with climatic, topographic and biotic factors. By considering a …
-
Assessing the Impacts of Agricultural Land Use on Wood Frog (Lithobates sylvaticus) Presence and Health in Central Saskatchewan
… neutrophil to lymphocyte (N:L) ratios), I used boosted regression trees, a relatively novel but growing modelling technique in the ecological sciences. Wood frogs were present in both grassland and cropland sites. eDNA was more successful at detecting wood frogs in wetlands compared to …
-
Applications of Probabilistic Machine Learning Models to Semiconductor Fabrication
… compares favourably to a number of standard regression approaches, including polynomial models, multivariate splines, gradient boosted regression trees, and a number of different deep learning architectures. While these four case studies each consider a unique application of probabilistic …
-
Investigating the movement and foraging ecology of reef manta rays (Mobula alfredi) in the Chagos Archipelago
Reef manta rays (Mobula alfredi) are large zooplanktivorous elasmobranchs of the family Mobulidae. The species is dispersed throughout the Indian and West Pacific Oceans in small, isolated populations that tend to have a limited home range. Individuals within a population regularly visit specific …
-
Temporal, Spatial, and Environmental Influences on the Demographics and Harvest Vulnerability of American Black Bears (Ursus americanus ) in Urban Habitats in New Jersey, Pennsylvania and West Virginia
… bears use urban and suburban habitats. We used boosted regression trees to create two predictive models of bear occurrence in urban and suburban habitats for (1) New Jersey and Pennsylvania, and (2) West Virginia. We separated West Virginia from New Jersey and Pennsylvania in the modeling …