Wavelet regression and additive models for irregularly spaced data

Authors: Asad Haris, Ali Shojaie, Noah Simon

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We complement our theoretical results with empirical studies comparing wave Mesh to existing methods. (Abstract) and 4 Numerical experiments (Section title)
Researcher Affiliation Academia Asad Haris Department of Biostatistics University of Washington Seattle, WA 98195 aharis@uw.edu Noah Simon Department of Biostatistics University of Washington Seattle, WA 98195 nrsimon@uw.edu Ali Shojaie Department of Biostatistics University of Washington Seattle, WA 98195 ashojaie@uw.edu
Pseudocode No The paper describes the algorithm in text (Section 2.3) and presents an iterative scheme (Equation 8), but does not provide a formally structured pseudocode or algorithm block.
Open Source Code No The R package wave Mesh, which implements our methodology, will soon be publicly available on Git Hub.
Open Datasets Yes We also analyze the motorcycle data studied by Silverman [1985] consisting of 133 head acceleration measurements in a simulated motorcycle accident taken at 94 unequally spaced time points. and For a real world data analysis, we consider the Boston housing data analyzed by Ravikumar et al. [2009].
Dataset Splits Yes Selection of tuning parameter for wave Mesh is done via 5-fold cross validation.
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies Yes The former two methods are implemented in the R package wavethres [Nason, 2016] and the latter is implemented in the adlift package [Nunes and Knight, 2017]. and Guy Nason. wavethresh: Wavelets Statistics and Transforms, 2016. URL https://CRAN. R-project.org/package=wavethresh. R package version 4.6.8.
Experiment Setup Yes We apply our proposal, wave Mesh, the interpolation proposal of Kovac and Silverman [2000] and isometric wavelet proposal of Sardy et al. [1999], for a sequence of 50 λ values linear on the log scale and select the λ value that minimizes the mean square error, MSE = n 1 f 0 b f 2 2.