Predicting Spatio-Temporal Propagation of Seasonal Influenza Using Variational Gaussian Process Regression

Authors: Ransalu Senanayake, Simon O'Callaghan, Fabio Ramos

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on two datasets of state-wide US weekly flu-counts consisting of 19,698 and 89,474 data points, ranging over several years, illustrate the robustness of the model as a tool for further epidemiological investigations. Experiments To demonstrate the predictive power of the method for the propagation of influenza, two datasets were used:
Researcher Affiliation Collaboration Ransalu Senanayake,1,2 Simon O Callaghan,2 and Fabio Ramos1,2 1School of Information Technologies, The University of Sydney, Australia 2National ICT Australia (NICTA)
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes 1Supplementary materials https://goo.gl/VCu VW3
Open Datasets Yes 1) Google flu trend (GFT): Recent studies show an increasing interest in using web query data to predict flu (Chakraborty et al. 2014). The US state-wide GFT (Ginsberg et al. 2009) data (flu count/population) of 402 weeks from 02/Dec/2007 to 09/Aug/2015 were used in the experiments. ... 2) 1972-2006 dataset (CDC): In order to analyze long-term trends, exact ILI counts (CDC ) of 1826 weeks from 1972 to 2006 (Viboud et al. 2006) were used.
Dataset Splits No The paper mentions 'training data' and 'test' data, but does not explicitly provide details about a separate 'validation' dataset split or its size/percentage for reproducibility.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts, or detailed computer specifications) used for running its experiments. It only vaguely refers to a 'standard PC'.
Software Dependencies No The paper does not provide specific software dependency details, such as library names with version numbers, used for the implementation or experiments.
Experiment Setup Yes in our experiments empirical rates of 10 2 and 10 5 were used as the learning rates of NGD and SGD respectively. Choosing only M = N/4 of data as inducing points average of MSE was found to be 0.0014 for GFT. M = N/10.