Global ETD Search

1	Performance Comparison of Public Bike Demand Predictions: The Impact of Weather and Air Pollution Min Namgung (9380318) 15 December 2020 (has links) Many metropolitan cities motivate people to exploit public bike-sharing programs as alternative transportation for many reasons. Due to its’ popularity, multiple types of research on optimizing public bike-sharing systems is conducted on city-level, neighborhood-level, station-level, or user-level to predict the public bike demand. Previously, the research on the public bike demand prediction primarily focused on discovering a relationship with weather as an external factor that possibly impacted the bike usage or analyzing the bike user trend in one aspect. This work hypothesizes two external factors that are likely to affect public bike demand: weather and air pollution. This study uses a public bike data set, daily temperature, precipitation data, and air condition data to discover the trend of bike usage using multiple machine learning techniques such as Decision Tree, Naïve Bayes, and Random Forest. After conducting the research, each algorithm’s output is evaluated with performance comparisons such as accuracy, precision, or sensitivity. As a result, Random Forest is an efficient classifier for the bike demand prediction by weather and precipitation, and Decision Tree performs best for the bike demand prediction by air pollutants. Also, the three class labelings in the daily bike demand has high specificity, and is easy to trace the trend of the public bike system. Applied Computer Science performance comparisons Public bike-sharing systems NYC bike trip data Chicago bike trip data Decision Tree Approach Research Random Forest models Naive Bayes models demand prediction air pollution impacts weather impacts Transportation Demand Management
2	Detecting and Measuring Corruption and Inefficiency in Infrastructure Projects Using Machine Learning and Data Analytics Seyedali Ghahari (11182092) 19 February 2022 (has links) Corruption is a social evil that resonates far and deep in societies, eroding trust in governance, weakening the rule of law, impairing economic development, and exacerbating poverty, social tension, and inequality. It is a multidimensional and complex societal malady that occurs in various forms and contexts. As such, any effort to combat corruption must be accompanied by a thorough examination of the attributes that might play a key role in exacerbating or mitigating corrupt environments. This dissertation identifies a number of attributes that influence corruption, using machine learning techniques, neural network analysis, and time series causal relationship analysis and aggregated data from 113 countries from 2007 to 2017. The results suggest that improvements in technological readiness, human development index, and e-governance index have the most profound impacts on corruption reduction. This dissertation discusses corruption at each phase of infrastructure systems development and engineering ethics that serve as a foundation for corruption mitigation. The dissertation then applies novel analytical efficiency measurement methods to measure infrastructure inefficiencies, and to rank infrastructure administrative jurisdictions at the state level. An efficiency frontier is developed using optimization and the highest performing jurisdictions are identified. The dissertation’s framework could serve as a starting point for governmental and non-governmental oversight agencies to study forms and contexts of corruption and inefficiencies, and to propose influential methods for reducing the instances. Moreover, the framework can help oversight agencies to promote the overall accountability of infrastructure agencies by establishing a clearer connection between infrastructure investment and performance, and by carrying out comparative assessments of infrastructure performance across the jurisdictions under their oversight or supervision. Policy Corruption Inefficiency Economics Econometrics Data analytics Time series analysis Machine learning Random forest models Artificial Neural Networks methods Principal component analysis (PCA) Cluster analysis Panel vector autoregression (PVAR) Impulse-response function
3	Digital Soil Mapping of the Purdue Agronomy Center for Research and Education Shams R Rahmani (8300103) 07 May 2020 (has links) This research work concentrate on developing digital soil maps to support field based plant phenotyping research. We have developed soil organic matter content (OM), cation exchange capacity (CEC), natural soil drainage class, and tile drainage line maps using topographic indices and aerial imagery. Various prediction models (universal kriging, cubist, random forest, C5.0, artificial neural network, and multinomial logistic regression) were used to estimate the soil properties of interest. Agricultural Land Management Agricultural Land Planning Sustainable Agricultural Development Agronomy Digital soil mapping (DSM) Organic matter content Cation exchange capacity Natural soil drainage classes geostatistical methods Cubist modeling Random Forest models Universal Kriging C5.0 or see5 modeling Artificial neural network Tile drainage lines Soil spatial variation Plant Phenotyping Aerial Imagery digital elevation model (DEM) terrain attributes Soil and Landscape Very poorly drained poorly drained somewhat poorly drained moderately well drained User's accuracy producer's accuracy kappa coefficient Multinomial logit regression Tile Probe topographic wetness index Topographic Position Index slope height cross sectional curvature Relative slope position channel network distance SSURGO soils data internal validation external validation

1

Page generated in 0.1152 seconds