The field of statistical learning theory has developed alternatives to induction. Instead of using all the available points to induce a model, the data, or usually a small subset of the data, can be used to estimate unknown properties of points to be tested (e.g., membership to a class). This idea leads to algorithms that use standard statistical tests to compute the confidence on the estimation. Using transduction, researchers have built transductive confidence machines which are able to estimate the unknown class of a point and attach confidence to the estimate, and also to determine outliers in a data set. WIREs Comp Stat 2011 3 216–220 DOI: 10.1002/wics.154
For further resources related to this article, please visit the WIREs website.
This article is a U.S. Government work, and as such, is in the public domain in the United States of America.