How do I filter out low occurrence nodes from a graph?

Question

I'm trying to make a graph using hashtags from Instagram. Each node is a hashtag and has edges to each hashtag it was paired with in a post. I want to filter out the nodes (hashtags) with a low number of occurrences.

I'm able to filter them out by manually setting a limit like 35 or 30, but I want to make it so this limit is calculated using some parameters from the graph.

Answer 1

You do not say how you want to adjust the limit according to which parameters of the graph.

Guess: you want to set the limit so that it eliminates some percentage of the nodes. e.g. you want to eliminate 10% of the nodes, those that have the lowest occurrence count.

• LOOP over every hashtag
  • Count occurences
  • save in array of pairs ( hashtag, #occurences )
• Sort pair array in ascending number of #occurences
• LOOP over sorted pair array
  • delete nodes with hashtag
  • IF required percentage of nodes have been deleted
    • STOP