Open prakhar987 opened 7 years ago
If a word pair is absent , is avoiding it when calculating spearman's correlation the best approach? I have seen people take mean vectors in place of missing words (mean of the entire vector space).
Is avoiding missing words the correct approach?
If a word pair is absent , is avoiding it when calculating spearman's correlation the best approach? I have seen people take mean vectors in place of missing words (mean of the entire vector space).
Is avoiding missing words the correct approach?