Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Divergence-from-randomness model
open-in-new
<p>In the field of <a href="/facts/Information_retrieval/MFSmpeNw">information retrieval</a>, divergence from randomness (DFR), is a <a href="/facts/Generalization/v0b1pwUx">generalization</a> of one of the very first models, Harter's <a href="https://link.springer.com/referenceworkentry/10.1007/978-0-387-39940-9_920">2-Poisson indexing-model</a>. It is one type of <a href="/facts/Probabilistic/fgYvMhwb">probabilistic</a> <a href="/facts/Model/Hj8KMwdD">model</a>. It is used to test the amount of <a href="/facts/Information/3LdrzXtJ">information</a> carried in <a href="/facts/Document/H4Hdxu91">documents</a>. The 2-Poisson model is based on the <a href="/facts/Hypothesis/gMABEJul">hypothesis</a> that the level of documents is related to a set of documents that contains words that occur in relatively greater extent than in the rest of the documents. It is not a 'model', but a framework for <a href="/facts/Weighting/HUltmeDW">weighting</a> terms using <a href="/facts/Probabilistic_method/AXKXrzrr">probabilistic methods</a>, and it has a special <a href="/facts/Relation_(mathematics)/HjblDaMy">relationship</a> for term <a href="/facts/Weighting/HUltmeDW">weighting</a> based on the notion of <a href="/facts/Elite/8ceFK8LD">elite</a> </p><p>Term weights are being treated as the standard of whether a specific word is in that set or not. Term weights are <a href="/facts/Computing/Kghtm9fX">computed</a> by measuring the <a href="/facts/Divergence_(statistics)/8qrn1djm">divergence</a> between a term distribution produced by a <a href="/facts/Stochastic_process/Ng1n1dnB">random process</a> and the actual term <a href="/facts/Distribution_(mathematics)/yfaJ4mUp">distribution</a>. </p><p>Divergence from <a href="/facts/Randomness/QJQBoEX5">randomness models</a> set up by instantiating the three main components of the framework: first selecting a basic randomness model, then applying the first normalization and at last normalizing the <a href="/facts/Term_Frequency/JDzc5Jnr">term frequencies</a>. The basic models are from the following tables. </p>