Content deleted Content added
→Super-online?: in machine learning |
Tag: |
||
(7 intermediate revisions by 5 users not shown) | |||
Line 1:
{{WikiProject
{{WikiProject Computing|importance=|auto=yes}}
}}
{{annual readership|scale=log}}
==Super-online?==
Is there a word for an online algorithm that has the additional property that partial results can be combined? That is, if I want to run an algorithm in parallel, I want to be able to combine partial results. For example, the sum of a set of numbers is the sum of the partial sums. [[User:BenFrantzDale|—Ben FrantzDale]] 01:00, 7 December 2006 (UTC)
Line 30 ⟶ 34:
Seems to describe the same class of algorithms. [[User:Qwertyus|Q<small>VVERTYVS</small>]] <small>([[User talk:Qwertyus|hm?]])</small> 09:59, 21 July 2013 (UTC)
:Then you don't understand what they are about, because they are very different. Online alforithms are about decision making in the face of uncertaintly, and making a sequence of decisions that would not necessarily be globally optimal if you knew the whole input at once but are as close to optimal as you can make with your limited knowledge. Streaming algorithms are about solving problems optimally (or close to optimally) when the whole input is given, but using sublinear memory to do it. —[[User:David Eppstein|David Eppstein]] ([[User talk:David Eppstein|talk]]) 16:31, 21 July 2013 (UTC)
== Misleading title for the "Definition" Section ==
I think the title of that section should be replaced, not sure to what title though. The content of the section does not give the definition of what is an online algorithm (it does provide a definition of the term "competitive ratio"), rather it discusses (not necessarily all main) properties of (not necessarily all) online algorithms.--[[User:Shay Zakov|Shay Zakov]] ([[User talk:Shay Zakov|talk]]) 08:08, 6 August 2020 (UTC)
|