artificial intelligence for students
14.9K views | +0 today
 
Scooped by Scott Turner
onto artificial intelligence for students
March 5, 2013 4:11 PM
Scoop.it!

Visualizing neural networks from the nnet package

Visualizing neural networks from the nnet package | artificial intelligence for students | Scoop.it
Neural networks have received a lot of attention for their abilities to ‘learn’ relationships among variables.
luiy's curator insight, March 14, 2013 7:58 AM

Neural networks have received a lot of attention for their abilities to ‘learn’ relationships among variables. They represent an innovative technique for model fitting that doesn’t rely on conventional assumptions necessary for standard models and they can also quite effectively handle multivariate response data. A neural network model is very similar to a non-linear regression model, with the exception that the former can handle an incredibly large amount of model parameters. For this reason, neural network models are said to have the ability to approximate any continuous function.

 

I’ve been dabbling with neural network models for my ‘research’ over the last few months. I’ll admit that I was drawn to the approach given the incredible amount of hype and statistical voodoo that is attributed to these models. After working with neural networks, I’ve found that they are often no better, and in some cases much worse, than standard statistical techniques for predicting relationships among variables. Regardless, the foundational theory of neural networks is pretty interesting, especially when you consider how computer science and informatics has improved our ability to create useful models.

artificial intelligence for students
artificial intelligence for students - postgraduate and undergraduate
Curated by Scott Turner