I am currently working on a classification algorithm. Each class is represented by a set of 3D vectors. The cardinality differs for each class. The order of the vectors in a set is completly random.
To assign a new measurement (again represented by a set for 3D vectors) I have to find the class with the minimum distance. However, i am not able to define a metric to measure the distances between the sets.
Any advise or hint for further research is welcome!
EDIT: Imagine the follwing example: class A is represented by a set with cardinality 10, class B is represented by a set with cardinality 5. My observation is represented by a set with 7 3D vectors. How can I calculate the distance between my observation and the two classes?
It seems like you are doing clustering? I am not sure what you mean by distance (you first say that a class has a distance, and then you say you want to "measure the distances between the sets"), but if you are referring to distances between points, then see this. Reading up on other introductions to clustering would probably be helpful.