Application of pairwise-correlations in Legal Risk Management

Investigate words that tend to co-occur within particular documents or particular chapters, even if they don’t occur next to each other. Consider the book “Pride and Prejudice” divided into 10-line sections.

Interpretation

• The graph shows words from Pride and Prejudice that were most correlated with ‘elizabeth’, ‘pounds’, ‘married’, and ‘pride’.
• It indicates how often they appear together in 10-line sections relative to how often they appear separately.
• Legal Risk Management Application: I think it could be a useful tool but can’t think of any specific examples. Any ideas?