Spatio-temporal modeling of the US college crime data
| Author(s) | Gezer, Fatih | |
| Date Accessioned | 2018-02-20T12:26:45Z | |
| Date Available | 2018-02-20T12:26:45Z | |
| Publication Date | 2017 | |
| SWORD Update | 2017-11-10T14:20:21Z | |
| Abstract | College crime is one of the most alarming social problems in the US today. To investigate important factors that are associated with college crime, we collected data from several publicly accessible sources and performed exploratory and statistical analyses. For the statistical analysis, Bayesian hierarchical modeling via Markov chain Monte Carlo and stepwise model selection procedures were applied to analyze such spatio-temporal data. We found the best models for California and Texas respectively in the sense that each model not only achieves a good balance between goodness-of-fit and interpretability but also satisfies spatial stationarity. A strong autoregressive effect was found for both states. The results additionally show that the proportion of undergraduate students and tuition are the most essential predictive factors that affect the college crime rate in California, while no strong factor is founded for Texas. | en_US |
| Advisor | Zhang, Xiaoke | |
| Degree | M.S. | |
| Department | University of Delaware, Department of Applied Economics and Statistics | |
| Unique Identifier | 1023626278 | |
| URL | http://udspace.udel.edu/handle/19716/23036 | |
| Language | en | |
| Publisher | University of Delaware | en_US |
| URI | https://search.proquest.com/docview/1972843837?accountid=10457 | |
| Keywords | Pure sciences | en_US |
| Keywords | Social sciences | en_US |
| Keywords | Education | en_US |
| Keywords | Autoregressive model | en_US |
| Keywords | Bayesian | en_US |
| Keywords | College crime | en_US |
| Keywords | Spatial stationarity | en_US |
| Keywords | Spatio-temporal modeling | en_US |
| Title | Spatio-temporal modeling of the US college crime data | en_US |
| Type | Thesis | en_US |