When can one consider a subgraph as a giant component of a network?

I am doing network resilience analysis of a co-occurrence network of words.

What I would like to understand is that, what is the minimum fraction of total nodes that must be present in a connected component(sub-graph) of a network for it to be considered as a giant com开发者_如何学运维ponent.

For example, in a network of 20,000 nodes if the maximum nodes a sub-graph contains is say 3, can it be considered a giant component?

As I understand, you are asking about the definition of the term "giant component".

It is a (qualitative) observation that if you add "enough" edges to a graph (especially in the case of random graphs), there will be a one component that contains the majority of the nodes. This is usually called the giant component.

This is a qualitative observation. There's no precise definition of "giant component" based on a fraction of nodes belonging to it. Only the observation that in a random graph of lots of nodes, there's a threshold connectivity around which the fraction of nodes belonging to the largest component will increase very sharply.

Is there a problem you are trying to solve or understand, or are you just asking about the definition of this term?