3
$\begingroup$

I'm currently working on a project where I'm supposed to compare the efficiency of SVM vs RVM, there seems to be a lot of information to be gathered about RVM whereas I find rather old documents about RVM. It's the first time I use this forum so I hope my question isn't off-topic but my questions are:

1) Which datasets are more suited for the two models and what are the upsides/downsides for each model? 2) As mentioned above, the articles I find about SVM are rather old, I'm currently curious about recent applications of SVM and to what and where they have been implemented.

Sources or just standard responses are highly appreciated. Thanks in advance

$\endgroup$
2
$\begingroup$

RVM is identical to SVM, but provides a probability distribution of scores.

RVM:

Pros:

  1. RVM is better than SVM in terms of accuracy.(depends on size of training data), more training data implies more accuracy in RVM. (argued and experimentally shown in the paper [1])

  2. Since it provides the probability distribution of score, not a point estimate. It opens up a variety of options to get point estimate out of Bayesian inference, like percentile, mean, median, average.

Cons:

  1. Since it uses EM algorithm for learning, it might get stuck at local minima.
  2. Large data is required for learning

SVM:

Pros:

  1. Standard SVM is a convex optimization problem(both hard and soft), therefore they always produce unique solution i.e. converge to global optima.

There is one very good paper about this. You can get more details here:

[1] (PDF) A comparative study of relevant vector machine and support vector machine in uncertainty analysis., accessed Jan 08 2019.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.