accessing a target comprising a target alphanumerical string that comprises one or more words;

determining one or more tokens from one or more documents, each token comprising a token alphanumerical string that comprises one or more words and appears at least once in the documents;

determining a token order among the tokens;

communicating the target as an original query to a search engine;

accessing one or more original search results returned by the search engine in response to the original query;

calculating an original value for each token based at least in part on a first number of appearances of the token in the documents and a second number of appearances of the token in the original search results;

generating a target vector representing the target, the target vector comprising the original values calculated for the tokens ordered according to the token order;

for each of a plurality of candidates comprising a candidate alphanumerical string that comprises one or more words:

generating a candidate vector representing the candidate; and

calculating a distance between the candidate vector and the target vector; and

selecting a candidate from the plurality of candidates that is related to the target based on a comparison among the distances between the candidate vectors and the target vector.