Applying Random Indexing to Structured Data to Find Contextually Similar Words