Here you can select which columns should be shown:

Functional similarity.

"Following (Turney, 2012), we define the functional similarity of two concepts to be the degree to which two concepts are observed with the same statements. The concepts 'coffee' and 'tea' for example are functionally similar, because they share a large set of statements, such as '[SO/STH] may drink [ ]', '[SO/STH] may sip [ ]' and '[SO/STH] may pour [ ]'.

We measure the similarity of two concepts or statements using the cosine distance (Bullinaria and Levy, 2007) over shared attributes in a vector space presentation. The closer the cosine distance is to 0, the more similar two concepts or statements are. We describe the process of calculating the cosine distance in greater detail in (Akbik et al., 2012)."

Applicative similarity.

"We define two statements to have a high applicative similarity if they are observed with similar nouns. For example, the statements '[ ] may cure [disease]' and '[SO/STH] may be administered [ ]' have a certain applicative similarity because they are observed with nouns such as 'medicine', 'dose' and 'draught'. A verbalization of this example would be 'things that cure diseases may often be administered to someone'.

We measure the similarity of two concepts or statements using the cosine distance (Bullinaria and Levy, 2007) over shared attributes in a vector space presentation. The closer the cosine distance is to 0, the more similar two concepts or statements are. We describe the process of calculating the cosine distance in greater detail in (Akbik et al., 2012)."