measure(ment),24-1-E99-1005,ak with five corpus-based variables : <term> co-occurrence frequency </term> of the <term> adjective-noun pair </term>
measure(ment),31-1-E99-1005,ak the <term> adjective-noun pair </term> , <term> noun frequency </term> , <term> conditional probability </term>
other,38-1-E99-1005,ak conditional probability </term> of the <term> noun </term> given the <term> adjective </term> ,
other,41-1-E99-1005,ak </term> of the <term> noun </term> given the <term> adjective </term> , the <term> log-likelihood ratio </term>
measure(ment),48-1-E99-1005,ak <term> log-likelihood ratio </term> , and <term> Resnik 's ( 1993 ) selectional association measure </term> . The highest correlation is obtained
measure(ment),7-2-E99-1005,ak highest correlation is obtained with the <term> co-occurrence frequency </term> , which points to the strongly lexicalist
other,20-2-E99-1005,ak lexicalist and collocational nature of <term> adjective-noun combinations </term> . This paper proposes an approach
hide detail