Using document access sequences to recommend customized information

WordSieve, a text analysis algorithm, uses a competitive-network-learning approach to learn topic-relevant keywords in real time with no predetermined corpus. You can use these keywords to form search engine queries to suggest relevant documents to the user.