What are the methods to find similarity between two document?
Answers
Answered by
2
hello mate,
With LSA you can fold documents into your semantic space as they are added; you do not need to recompute the SVD every time. Obviously, over time, you will need to recompute it at some point (say when some proportion of the documents in yoru semantic space were 'folded in'). To compare two documents using the cosine similarity measurement you simple scale both document vectors by their corresponding singular values and compute cosine as per usual (dot product over product of magnitudes).
hope it help you!
With LSA you can fold documents into your semantic space as they are added; you do not need to recompute the SVD every time. Obviously, over time, you will need to recompute it at some point (say when some proportion of the documents in yoru semantic space were 'folded in'). To compare two documents using the cosine similarity measurement you simple scale both document vectors by their corresponding singular values and compute cosine as per usual (dot product over product of magnitudes).
hope it help you!
Similar questions