Efficiently build a graph of words with given Hamming distance

Question

I want to build a graph from a list of words with Hamming distance of (say) 1, or to put it differently, two words are connected if they only differ from one letter (lo

Answer 1

Ternary Search Trie supports Near-Neighbor Searching pretty well.

If your dictionary is stored as TST then, I believe, average complexity of lookups while building your graph would be close to O(N*log(N)) on real world word dictionaries.

And check Efficient auto-complete with a ternary search tree article.