Understanding Dominant Factors for Precipitation over the Great Lakes Region

Authors: Soumyadeep Chatterjee, Stefan Liess, Arindam Banerjee, Vipin Kumar

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments show that global climate indices, computed at different temporal lags, offer predictive information for winter precipitation.
Researcher Affiliation Academia 1Department of Computer Science and Engineering., 2Department of Soil, Water, and Climate University of Minnesota, Twin Cities Minneapolis, MN 55455 {chat0129, liess}@umn.edu, {banerjee, kumar}@cs.umn.edu
Pseudocode No The paper does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper provides a link to 'Supplementary material' (http://www-users.cs.umn.edu/chatter/papers/15/supplement.pdf), but it does not explicitly state that this material includes the open-source code for the methodology described in the paper.
Open Datasets Yes We compiled datasets from two sources: (1) United States Historical Climatological Network (USHCN) (Menne, Williams Jr, and Vose 2010), and (2) North American Regional Reanalysis (NARR) (Mesinger, Di Mego, and others 2006).
Dataset Splits Yes We divided the data into two sets. The first, comprising of 22 years data, was used for finding dominant factors. The second set, with the remaining 10 years data, was used to test predictive performance. ... We conducted leave-one-year-out cross-validation on held out test set described earlier.
Hardware Specification No The paper acknowledges technical support from the 'University of Minnesota Supercomputing Institute (MSI)', but it does not provide specific hardware details (e.g., CPU/GPU models, memory, or number of machines) used for running the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., Python 3.8, PyTorch 1.9, CPLEX 12.4). It describes methodologies like LASSO and Ordinary Least Squares regression.
Experiment Setup Yes For choosing the regularization parameter λ, we selected 2% of the training set as a validation set and selected λ that provides the smallest prediction mean square error (MSE) on this validation test.