How can we interprete the results generated by SVM? I am using SVM for classification purpose. I got results but I am not understanding how to interpret its results and also how can I know the contribution of each independent variable in the prediction accuracy of SVM? 
Please share any link of a case-study using R.
 A: If you want to understand your results you can simply compute the training error and the test error. You have always to check that you are not overfitting, and in order to see if it's the case, you can use the cross validation strategy, try to plot the learning curves and see the difference between the Validation error and the training error. 
Usually the coefficients of each independent variable tell us the importance of each feature. If you have a very small (close to 0) coefficient associated to an independent variable it's probably due to the fact that the model is not using that variable so much. It's not easy to interpret the weights, try to convince yourself that weights define simply an hyperplane for separating data. Now try to think about the 2D case. If you have one of the two coefficient that is close to zero you have a plane that is orthogonal to one of the two axes. So probably your data can be separated by this type of hyperplane.
Another thing you can try to do in order to obtain better results in the case you are working in high dimensional features space is using kernel methods and try to find the appropriate kernel, there are several ways in order to find the best kernel, for example the Mercer’s condition. 
I never programmed in R so i can't give you an example of code.  
A: SVM is a classification algorithm that relies on optimization only. It does not assume a probabilistic model. You can use it for prediction, but not really for inference. FraMan explanation might give some intuition, but I'm not sure how it generalizes to different kernels than the linear one, and I'm not 100% sure it holds for the linear as well.
