jackHI

Context

Histograms are a commonly used data representation method that can be thought of as a (non-parametric) estimate of the probability distribution function of the data-generating process. They are also frequently used to determine summary quantities of this function. The values of such summary quantities, however, depend strongly on the way the bins are chosen.

Purpose

jackHI is a software suite that contains functions that help a user generate a histogram based on a non-arbitrary, data-based optimal choice of bin width. The idea is that one samples the underlying probability distribution function sufficiently to capture the data features, whereas fine details due to `random noise’ (caused by empty too-finely sampled bins) are ignored. This sampling is effected by a “jackknife” (or cross-validation) method (see Hogg (2008))¹.

Applications

Work in progress.

Citations/using this code

Work in progress.