Question $101$ Frequency and positive wise mutual information are two possible measures of association between two words which might be used in distributional representations of words. Positive wise mutual information is generally considered as superior to frequency because it cannot be negative $0$ a co$-$occurring word is only considered as a feature if it occurs more than a threshold number of times if two words co$-$occur together more than we would expect if words occurred independently, then this will have more weight in the similarity calculation if two words co$-$occur together more than we would expect if words occurred independently, then this will have less weight in the similarity calculation it is mor UN$20$ e difficult to calculate