Talk:Decision boundary

This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Robotics Mid‑importance This article is within the scope of WikiProject Robotics, a collaborative effort to improve the coverage of Robotics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.RoboticsWikipedia:WikiProject RoboticsTemplate:WikiProject RoboticsRobotics articles Mid This article has been rated as Mid-importance on the project's importance scale. Statistics Low‑importance This article is within the scope of WikiProject Statistics, a collaborative effort to improve the coverage of statistics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.StatisticsWikipedia:WikiProject StatisticsTemplate:WikiProject StatisticsStatistics articles Low This article has been rated as Low-importance on the importance scale.

I wrote this article to replace what was basically a substub, but I don't know much about this topic, so I've tagged it for expert review. (I'd never even heard of support vector machines until now; I just mentioned them because they were mentioned in the original version of the article.) I also don't know what a decision space is, or how it relates to the concept of a decision boundary, but perhaps it ought to be merged somewhere (here, for example)? —User:Caesura^(t) 19:57, 28 November 2005 (UTC)[reply]

As far as I can tell, a decision space is just a 3 dimensional hyperplane belonging to a 4 dimensional hyperspace, which then seperates 4 dimensional objects in two groups. I only study this field right now, so it may be completly off.

I cite the http://en.wikipedia.org/wiki/Universal_approximation_theorem

The article reads "If it has one hidden layer, then it can learn problems with convex decision boundaries (and some concave decision boundaries). The network can learn more complex problems if it has two or more hidden layers." This is not true. A feedforward neural network with one hidden layer and one output layer can approximate anything. More layers may be more efficient, more elegant, or otherwise desirable, but not strictly needed. — Preceding unsigned comment added by 70.162.89.24 (talk) 23:02, 26 February 2014 (UTC)[reply]