I've completed project 0 - degrees using the general DFS / BFS algorithm. Needless to say it is rather slow. I've been thinking but haven't yet found a good heuristic to speed up the search when removing nodes from the frontier.

In the lecture example, it's quite obvious that we can use some kind of distance between 2 nodes, since the problem is literally about distance. With the imdb data, we have the birth year (for people) and year of release (for movies), but I can't think of any way how those info could be used to come up with some "distance" measure between 2 people. Anyone has any suggestion for a good heuristic to use for this problem?

Many thanks.

  • Hey there, yeah, my one is pretty slow too. Check out oracleofbacon.org/how.php They have a couple of suggestions there, like assigning a Kevin Bacon number to each actor before executing the searches. But as to implementation, that's still a bridge too far for me at the moment. – cormacl Mar 10 at 20:15

You must log in to answer this question.

Browse other questions tagged .