Dear Muthukumaran,

SVM, ANN, and k-NN are considered as state-of-the-art methods for weld detection and classification. A relevant journal article is given in the link below:

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.165.6726&rep=rep1&type=pdf

The title of the above given article is:

"Multiclass defect detection and classification in weld radiographic images using geometric and texture features".

Best of luck,

Hi,

You can find a lot of machine learning systems for data classification. But, it is worthy of consideration that we can categorize all of them into two main groups of local and global classification systems. Based on each dataset, features vector and size of dataset a local or global classifier can make  a better model for final classification. For datasets with a huge number of cases if you work on ANN methods, it is possible for you to have feature selection of a long feature vector and classification simultaneously. Hence, you can make your final feature vector efficient enough  and results are more reliable but not fast like KNN.

In a nutshell, for day to day research KNN and SVM are more conventional, for small datasets, it would be logical if you train and test one local and one global classifier on your data to find out which one makes better results. KNN and SVM are kind of state of the arts in this area.

KNN is an instance-based “lazy” machine learning method to build an optimal classification function locally.  and  SVM is an “eager” machine learning method, which is trained using the entire training samples to build a global model of fitting the training data.

Therefore, I recommend you to make model for your data by both of them to find out which one is better.

I guess you can try to evaluate deep learning classification that very popular and effective.

The important thing about research is to analyze all of the classifiers on your data.

Then you should use the one which give min loss and best validation accuracy on your data, it may be ANN, RNN, SVM, and KNN etc.

First of all, you need to identify your problem. It depends upon what kind of data you have and what your desired task is.

If you are Predicting Category:

You have Labeled DataYou need to follow Classification Approach and its algorithms

You don't have Labeled DataYou need to go for Clustering Approach

If you are NOT Predicting Quantity:

You need to go for Regression Approach

Otherwise:

You can go for Dimensionality Reduction Approach

There are different algorithms within each approach mentioned above. The choice of a particular algorithm depends upon the size of the dataset.

Things you might consider in choosing which algorithm to use would include:

1. Do you need to train incrementally (as opposed to batched)? If you need to update your classifier with new data frequently (or you have tons of data), you'll probably want to use Bayesian. Neural nets and SVM need to work on the training data in one go.

2. Is your data composed of categorical only? Or numeric only? Or both? I think Bayesian works best with categorical/binomial data. Decision trees can't predict numerical values.

3. Do you or your audience need to understand how the classifier works? Use Bayesian or Decision Trees, since these can be easily explained to most people. Neural networks and SVM are "black boxes" in the sense that you can't really see how they are classifying data.

4. How much "classification speed" do you need? SVM's are fast when it comes to classifying since they only need to determine which side of the "line" your data is on. Decision trees can be slow especially when they're complex (e.g. lots of branches).

Complexity. Neural nets and SVMs can handle complex non-linear classification.

Hope that helped.

Read this:

https://nlp.stanford.edu/IR-book/html/htmledition/choosing-what-kind-of-classifier-to-use-1.html

Dear Muthukumaran Malarvel,

The suitability of choosing the right classifier is depending on the types of data that you have. There are classifiers which are more suitable to use using qualitative set of data and vice versa. Hope this can help...

Dear Muthukumaran,

The selection of classifiers depend of many factor and usually is very difficult choose a one classifier. Some parameters as the type of data, complexity of classifier, accuracy, real time response and classifier' structure are essential for a important selection. First, I recommended search review papers based and classifiers and your application and later work with some classifier to evaluate your performance.

Best regards

I agree with Mr. Akay''s opinion. The selection of classifier depends on the type and size of selected features or data. I also agree with opinion given by Mortaza. It is better that you evaluate classifiers and then, make comparison on their performance.

Dear Muthukumaran

I suggest using SVM classifier because it is considered as state-of-art classification approach and you can to modify it a little bit to obtain contributions to your work.

Regards,

Best classifier depends on the your data type and the attributes you choose. you can try bayesian classifiers and SVM. Select the one with best accuracy

Classifier selection depends on the type,  size and characteristics of the input data or features. In my opinion, ANN, k-NN and SVM is among the top and robust classifiers.

If you have a limited and little dataset and your data is supervised, then base on machine learning theory it is better to use a classifier with high bias like Naive Bayes, although this effect is not necessarily observed in practice with regularized models over textual data. At any rate, a very low bias model like a nearest neighbor model is probably counter indicated. Regardless, the quality of the model will be adversely affected by the limited training data.

If you have tons of data in your dataset, then the choice of classifier probably has little effect on your results and the best selection may be unclear. It may be best to choose a classifier based on the scalability of training or even run-time efficiency. To get to this point, you need to have huge amounts of data.

have a look at:

https://nlp.stanford.edu/IR-book/html/htmledition/choosing-what-kind-of-classifier-to-use-1.html

you can also have a look at this link:

http://blog.echen.me/2011/04/27/choosing-a-machine-learning-classifier/

However, you can also get help from some machine learning softwares like Weka which can give you Tens of classifiers and interesting results to help you to make a good decision.

it would be better to run multiple of classifications algorithms and take a look at their evaluation results and make sure that random forests is one of them.

Good luck