I hope I posted this thread in the correct subforum.

Im trying to compare graphs basically and wonder what would be the best mathematical approach to a problem like this.

I have a graph that works as my key-graph. Then I have a set of candidate-graphs which I want to evaluate. They are evaluated so that a candidate-graph is considered better if it is more similar to the key-graph ( more correlated in some way). The key-graph have a fixed amount of points so between these points there is straight lines, so it cannot be any exponential appearance between the points in the graph. The problem is that the candidate-graphs may have more or less of these points so I can't correlate with the help of these points because there isnt the same amount of them in each graph.

Is it possible to use some correlation-approach to this problem? Or anyone have any suggesion on another way to be able to evaluate the candidate-graphs so that a more "similar" graph is considered better?