We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Testing for Feature Relevance: The HARVEST Algorithm

Abstract: Feature selection with high-dimensional data and a very small proportion of relevant features poses a severe challenge to standard statistical methods. We have developed a new approach (HARVEST) that is straightforward to apply, albeit somewhat computer-intensive. This algorithm can be used to pre-screen a large number of features to identify those that are potentially useful. The basic idea is to evaluate each feature in the context of many random subsets of other features. HARVEST is predicated on the assumption that an irrelevant feature can add no real predictive value, regardless of which other features are included in the subset. Motivated by this idea, we have derived a simple statistical test for feature relevance. Empirical analyses and simulations produced so far indicate that the HARVEST algorithm is highly effective in predictive analytics, both in science and business.
Comments: 22 pages, 2 tables
Subjects: Machine Learning (stat.ML)
Cite as: arXiv:1710.00210 [stat.ML]
  (or arXiv:1710.00210v2 [stat.ML] for this version)

Submission history

From: Herbert Weisberg [view email]
[v1] Sat, 30 Sep 2017 14:53:51 GMT (309kb)
[v2] Tue, 27 Feb 2018 21:44:12 GMT (378kb)

Link back to: arXiv, form interface, contact.