What are advantages of Artificial Neural Networks over Support Vector Machines?

Question

I know there are many advantages that SVMs have over ANN since they cover two main disadvantages of ANN:

1. ANN covers only the local minima and not global minima
2. ANN sometime considers noise as a part of a pattern

But I want to know” What specific advantage does ANN has over SVM which makes ANNs usable ?”

Answer 1

@Anisha, Following are the advantages of Artificial Neutral networks over SVM-

l. ANNs can produce any number of outputs while SVMs have only one output.

2. An n-ary classifier with a neural network can be trained in one go while n-ary classifier with SVMs trains each of them one by one.

3. SVMs are isolated systems whereas, ANNs are one whole.

4. ANNs are parametric models while SVMs are non-parametric models.

Ex-

If you had to measure a person's hormone level as a function of other physiological factors like heart rate etc, then it is preferable to use artificial neural network regression over support vector machine regression since those factors are inter-related.

Hope this answer helps.

Answer 2

One benefit that the ANN models have over SVMs is that their size is fixed: they are parametric models, while SVMs are non-parametric models.

Depending on the number of features, plus bias parameters, ANN has a bunch of hidden layers with sizes h1 to hn and those help in making up the model. By contrast, an SVM (at least a kernelized one) consists of a set of support vectors, selected from the training set, with a weight for each. SVM has classifiers with tens of thousands of support vectors, each having hundreds of thousands of features. And those large number of features are not known to us thus no use of those features.

One more advantage of artificial neural networks over support vector machines is that artificial neural networks may have any number of outputs, while support vector machines have only one. Thus, ANN can train the models in one go and SVM has to train one by one.

Watch this video to learn about Neural Networks: