Spatial analysis made easy with linear regression and kernels

• Published in: Epidemics Vol. 29; p. 100362
• Main Authors: Milton, Philip, Coupland, Helen, Giorgi, Emanuele, Bhatt, Samir
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.12.2019
• Subjects: Algorithms; Fourier Analysis; Humans; Infectious Disease; Internal Medicine; Kernel approximation; Kernel methods; Linear Models; Random Fourier features; Regression; Spatial Analysis; Regression; Random Fourier features; Kernel methods; Kernel approximation
• DOI: 10.1016/j.epidem.2019.100362
• ISSN: 1755-4365; 1878-0067

•Many kernel methods are only suitable for small/medium sized spatial problems.•Random Fourier features speeds up kernel method with a minimal drop in accuracy.•This speedup lets us efficiently work with large spatial problems.•They can be added into many common spatial methods with only a few lines of code. Kernel methods are a popular technique for extending linear models to handle non-linear spatial problems via a mapping to an implicit, high-dimensional feature space. While kernel methods are computationally cheaper than an explicit feature mapping, they are still subject to cubic cost on the number of points. Given only a few thousand locations, this computational cost rapidly outstrips the currently available computational power. This paper aims to provide an overview of kernel methods from first-principals (with a focus on ridge regression) and progress to a review of random Fourier features (RFF), a method that enables the scaling of kernel methods to big datasets. We show how the RFF method is capable of approximating the full kernel matrix, providing a significant computational speed-up for a negligible cost to accuracy and can be incorporated into many existing spatial methods using only a few lines of code. We give an example of the implementation of RFFs on a simulated spatial data set to illustrate these properties. Lastly, we summarise the main issues with RFFs and highlight some of the advanced techniques aimed at alleviating them. At each stage, the associated R code is provided.