The performance depends on the data sets being used. Please test with the classifiers first, and then interpret.

On my experiments SVM has better accuracy tan ANN in Natural Language Processing.

The classification model is not the most important part, if you have a good training dataset you can obtain good accuary. (sometimes the same accuracy).

If you want to increase accuracy ,you can try with 'Variable and Feature Selection'. Maybe this introduction paper Will be useful https://dl.acm.org/citation.cfm?id=944968

Dear Hamid,

In what field you want applied ANN or others similar technique? You must answer on this question before choosing a best intelligent technique. The choose will depend of application arry.

Best regards

The performance depends on the data sets being used. Please test with the classifiers first, and then interpret.

I know, It is very difficult to know which soft computing methods and training algorithm will be the fastest for a given problem. It depends on many factors, including the complexity of the problem, the number of data points in the training set, the number of weights and biases in the network, the error goal, and whether the network is being used for pattern recognition (discriminant analysis) or function approximation (regression). 1"

Function approximation (nonlinear regression) is the field that I would like to learn more. If datasets are available which soft computing methods could show better performance? Though one method should be chosen. How this selection done. Could propose a rational algorithm for this choice.

1-link:https://www.mathworks.com/help/nnet/ug/choose-a-multilayer-neural-network-training-function.html

Dear Hamid,

I think in this case the Neuro-fuzzy based system is able to approximate the non-linear function with more precision. For more assurance and credibility i suggest you to see links and attached files in subject.

- Application of non-linear regression analysis and artificial intelligence ...

www.uni-stuttgart.de/igs/content/.../314_AS_TUST58_2016.pdf

- A hybrid neural networks-fuzzy logic-genetic algorithm for grade ...

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4268588/

- Multilayer Perceptron for Robust Nonlinear Interval Regression ...

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4106049/

- Time Series Seasonal Analysis Based on Fuzzy Transforms - MDPI

www.mdpi.com/2073-8994/9/11/281/pdf

In my experience (clinical)

ANN performs better on non linear problems like (absorbance-photometry).

In clinical predictions, they seem to perform poorly against Random Forests.

SVM is definitely something nice.

Feature selection is nice.

Usually ANN are blackboxes, we can still extract knowledge from it.

In Medicine, using your own knowledge for the feature selection peforms actually better than any other feature selection.

In conclusion, please test all of the machine learning methods using a validation. Then chose the most efficient.

There are couple of packages in R such as carret who can help you with feature selection.

RBF is ANN too! ANFIS combines ANN with Fuzzy logic, etc

Generally SVM and ANN are good for classification tasks.

ANN, SVR and ANFIS are good for function approximation.

My project is to predict life cycle. I have done it in mathematially and I have done in anfis I got answer but is in like if answer is 45886 I am getting like 45000

The performance depends on the size of data sets being used. For example according to my experience in small data sets (500 data) ANFIS has better accuracy than ANN.

Article Intelligent Decision Support System for Industrial Site Clas...

Based on the theory, all the method have advantage and disadvantage. But in the practical we can not judge what method is the best. It depends your case and you should be tested.