[R] Evaluating statistical models and describing coefficients for non-parametric models

Jeff Newmiller jdnewmil at dcn.davis.ca.us
Tue May 17 17:43:12 CEST 2016

If the documentation for those models does not tell you how, then you should ask yourself whether what you are trying to do makes sense statistically. The question of what makes sense statistically is not a good fit for this forum, which is about R.

By the way, root mean square errors are easy to calculate... the hard part is knowing whether doing so is informative per above. 
Sent from my phone. Please excuse my brevity.

On May 17, 2016 3:56:56 AM PDT, Muhammad Bilal <Muhammad2.Bilal at live.uwe.ac.uk> wrote:
>Hi All,
>I'm using number of models such as lm(), tree, randomForest, svm, and
>nnet for predicting the delays in projects. Also, I computed the sum of
>squared error for all these models for comparison purposes. However, I
>want to use other related evaluation criteria such as root mean sum of
>square error (RMSE) and R Squared for evaluation of these models.
>My question is that is it possible to compute these criteria (RMSE or
>R2) for all above-mentioned statistical models.
>Second, for the lm() we can see the co-efficient values by checking
>model summary. Is it possible to see the co-efficient for other models
>such as SVM and neural network?
>Thanks in advance for the help and support.
>Many Thanks and
>Kind Regards
>Muhammad Bilal
>Research Fellow and Doctoral Researcher,
>Bristol Enterprise, Research, and Innovation Centre (BERIC),
>University of the West of England (UWE),
>Frenchay Campus,
>BS16 1QY
>muhammad2.bilal at live.uwe.ac.uk<mailto:olugbenga2.akinade at live.uwe.ac.uk>
>	[[alternative HTML version deleted]]
>R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>PLEASE do read the posting guide
>and provide commented, minimal, self-contained, reproducible code.

	[[alternative HTML version deleted]]

More information about the R-help mailing list