Posts

Suppose you have some data and you want to build a model to predict future observations. Data splitting means dividing your data into two, not necessarily equal, parts. One part is used to build the model and the other half to evaluate it in some way.

In What’s wrong with simultaneous confidence bands, I discussed the deficiencies of standard confidence bands, but how can we do better? Suppose we have a Gaussian process for the curve with some mean and a covariance kernel:

Sky surveys provide a wide-angle view of the universe. Instead of focusing on a single star or galaxy, they provide a snapshot of a wider portion of the sky. Their primary purpose is to search for asteroids, so they take a sequence of images over a short timespan: minutes or hours.

Here are some 95% confidence bands for a fitted curve shown on the same data. The first uses spline smoothing from the mgcv package while the second uses a loess smoother from the ggplot2 package.

While looking at tables of contents for journals and CVs for job applications, it struck me that there are a lot of multi-author papers now. Just to check that I was not imagining it, I downloaded the tables of contents for JASA and JRSS-B for the years 2014, 2004, 1994, 1984 and 1974.

Consider a regression dataset with a response and several predictors. You want a single plot showing the response plotted against each of the predictors. You could use the pairs() but that also shows plots between the predictors.

There are a very large number of books, articles and university courses with “Generalised Linear Models” (GLM) in the title. I even wrote such a book myself. But is the GLM paradigm really that useful?