Kurtoses

Part of speech: noun

Definitions

  1. The plural form denoting statistical measures that describe the peakedness or flatness of a frequency distribution curve compared to a normal distribution ; expressions used in statistics to quantify the extent of outlier presence and tail thickness in datasets ; numerical indicators employed to evaluate the sharpness or flatness of a distribution's shape relative to a Gaussian curve
  2. Terms representing values that assess how much a data distribution's tails differ from those of a normal distribution ; statistical descriptors showing whether data have heavy or light tails and how extreme the peaks are ; quantifications used to measure deviations in height and tail thickness of probability distributions compared to a bell curve
  3. Concepts in statistics referring to moments that characterize the extremity and shape of a dataset's distribution ; plural labels for measures that help identify if data have outliers by indicating tail weight and peak sharpness ; tools to analyze how data spreads or clusters focusing on distribution height and tail heaviness relative to standard curves

Etymology: Karl Pearson, the English mathematician, biometrician and eugenicist, coined the term "kurtosis" around 1895 and published it in 1899 in a paper on the Hypergeometrical Series in the Philosophical Magazine. He took it from Ancient Greek "kúrtōsis", meaning "a bulging, convexity", which itself came from "kurtós", "bulging, convex". Pearson needed a precise name for a property he was measuring, and he reached back to a Greek word describing a curved, swelling shape. The Greek image fits the measure well. Kurtosis describes how peaked a distribution is and how heavy its tails run, so a sharply bulging curve scores high and a flatter one scores low. The "-osis" ending is a common Greek noun suffix, used in medicine and science for conditions or states, which is why the coinage sounds technical to English ears. The plural "kurtoses" is simply the Greek-style plural of "kurtosis", the form statisticians use when they speak of several such measures at once.

Synonyms: kurtosis values, peakedness measures

Antonyms: platykurtosis, mesokurtosis