Thank you for the response sir. I got the idea that shifting means replacing the sample by a weighted linear combination of the signal samples at its neighbors. But I could not get the intuitive meaning behind this? For example, in Time series shifting a sample by 1 unit means replacing the sample by its previous sample. From this we go for left shift or right shift. Like wise , here what do we conclude when we say that a graph signal has shifted twice or thrice?
I hope I am not skewed : Coloring nodes is to characterize them as f(x1,x2....) . Now when you shift, based on new set of parameters they can be used to check say their equivalency .