2017-05-19: Basis functions paper discussion

[mbjoseph]

This friday we'll dive into some machine learning, discussing a recent paper on basis functions (link to pdf here).

Hefley, T.J., K.M. Broms, B.M. Brost, F.E. Buderman, S. Kay, H.R. Scharf, J.R. Tipton, P.J., Williams, M.B. Hooten (2017) The basis function approach for modeling autocorrelation in ecological data. Ecology 98: 632-646.

Basis functions provide the means for fitting smooth curves such as splines and Generalized Additive Models, and also incorporate linear regression, ANOVA, and many more standard approaches as special cases. Increasingly, spatial, temporal, and spatiotemporal approaches for modeling big data are employing new types of basis expansions to manage the dimensionality of parameter space. But this Friday, we'll start from the basics ;)

When: Friday 2017-05-19 2 PM Where: SEEC S340

[mbjoseph]

Here's a quick demo of using basis functions to predict maximum temperature in Boulder, CO: https://gist.github.com/mbjoseph/99cc32076bc92e0c273259d5b6c74b7a

And here is a link to interactive figures from the manuscript: https://hootenlab.shinyapps.io/hooten_lab_manuscript_figures/