Atomic Spatial Processes
Authors: Sean Jewell, Neil Spencer, Alexandre Bouchard-Côté
ICML 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The framework s superiority over both empirical grid-based models and Dirichlet process mixture models is demonstrated by fitting, interpreting, and comparing models of graffiti prevalence for both downtown Vancouver and Manhattan. This section illustrates two applications of our ASP model and inference strategy using graffiti data from downtown Vancouver and Manhattan. |
| Researcher Affiliation | Academia | Sean Jewell SEAN.JEWELL@STAT.UBC.CA Neil Spencer NEIL.SPENCER@STAT.UBC.CA Alexandre Bouchard-Cˆot e BOUCHARD@STAT.UBC.CA Department of Statistics, University of British Columbia, Vancouver, BC V6T 1Z4, Canada |
| Pseudocode | No | The paper describes the steps of the MCMC algorithm in prose, but no structured pseudocode or algorithm block (e.g., Algorithm 1, or a formatted code-like block) is provided. |
| Open Source Code | No | The paper does not provide any statement or link indicating the availability of open-source code for the described methodology. |
| Open Datasets | Yes | The cities of Vancouver and New York City each maintain records of all graffiti sites identified by city staff. The data is made publicly available through the Vancouver Open Data Catalog and NYC Open Data database.2 2Vancouver Open Data Catalog: http://data.vancouver.ca/datacatalogue/graffitiSites.htm. NYC Open Data database: https://data.cityofnewyork.us/City-Government/DSNY-Graffiti-Information/gpwd-npar |
| Dataset Splits | Yes | Performance is judged through a held-out analysis of 10% of the observed pieces of graffiti, conducting posterior inference using the reduced data, and calculating predictive likelihoods of the omitted data. For the DPM-empirical mixture, an additional 10% of the training data is held-out to cross validate the mixing proportions. |
| Hardware Specification | No | The paper states 'Computing was supported by West Grid.' but does not provide specific hardware details such as GPU or CPU models, processor types, or memory used for experiments. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers. |
| Experiment Setup | Yes | The inference of µ, the normalized version of µ, for Figure 1b was based on an MCMC strategy (outlined in Section 4) consisting of 20,000 iterations, with resampling of the hyperparameters α0, απ 0, c, κ, and ν. Starting values for the hyperparameters can be chosen by considering the mixing behaviour for varying short runs. We use random walk Gaussian kernels as the proposal for each Metropolis Hastings step. Each kernel is centered at the current value of the selected hyperparameter with a fixed variance specific to each hyperparameter. Short runs are used to calibrate the variance of each hyperparameter to ensure proper mixing. |