pyGPs -- A Python Library for Gaussian Process Regression and Classification
Authors: Marion Neumann, Shan Huang, Daniel E. Marthaler, Kristian Kersting
JMLR 2015 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We introduce py GPs, an object-oriented implementation of Gaussian processes (gps) for machine learning. The library provides a wide range of functionalities reaching from simple gp speciļ¬cation via mean and covariance and gp inference to more complex implementations of hyperparameter optimization, sparse approximations, and graph based learning. [...] py GPs has an object-oriented structure and it additionally supports useful routines for the practical use of gps, such as cross validation functionalities for evaluation as well as basic routines for iterative restarts for gp hyperparameter optimization. [...] The implemented evaluation measures are root mean squared error (RMSE), accuracy (ACC), precision and recall (Prec, Recall) and the negative log predictive density (NLPD) to evaluate the quality of the whole predictive gp model. |
| Researcher Affiliation | Collaboration | Marion Neumann EMAIL Department of Computer Science and Engineering, Washington University, St. Louis, MO 63130, United States; Shan Huang EMAIL Fraunhofer IAIS, 53757 Sankt Augustin, Germany; Daniel E. Marthaler EMAIL Sproutling, San Francisco, CA 94111, United States; Kristian Kersting EMAIL Department of Computer Science, TU Dortmund University 44221 Dortmund, Germany |
| Pseudocode | No | The paper provides Python code snippets to illustrate the library's usage (e.g., '1 model = py GPs.GPR()'), but these are executable code examples, not abstract pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | Yes | py GPs is released under the Free BSD license and it can be downloaded from http:// mloss.org/software/view/509/ or https://github.com/marionmari/py GPs. |
| Open Datasets | No | The paper introduces a software library and describes its functionalities. While it mentions the library's capabilities for various ML tasks (regression, classification) and evaluation (cross-validation, metrics), it does not present specific experiments using datasets nor provides access information for any datasets used within the paper itself. |
| Dataset Splits | No | The paper describes a software library for Gaussian processes and its evaluation features like k-fold cross-validation. However, it does not present results from specific experiments on datasets in the paper, and therefore does not specify any dataset splits for reproducibility. |
| Hardware Specification | No | The paper describes a software library and its implementation details. It does not contain any specific information about hardware used for development or evaluation within the paper's scope. |
| Software Dependencies | Yes | py GPs requires Python 2.6 or 2.7 (www.python.org) and the numpy (www.numpy.org), scipy (www. scipy.org), and Matplotlib (www.matplotlib.org/) packages. |
| Experiment Setup | No | The paper describes the functionalities of the pyGPs library, including default settings for GP regression (e.g., zero mean, RBF kernel, Gaussian likelihood, exact inference, 'minimize' optimizer). However, it does not provide specific hyperparameter values or training schedules for any experiments conducted by the authors and presented in this paper; it rather outlines the library's capabilities. |