Lexical Chaining is an important technique in natural language processing, many of Lexalytics algorithms rely heavily on it.
Lexical Chaining relates sentences via thesaurally-related nouns. Consider the following:
I like beer. Miller just launched a new pilsner. But, because I’m a beer snob, I’m only going to drink pretentious Belgian ale.
Those 3 sentences are related through
Even if those sentences are not adjacent to each other in the text, they are lexically related to each other and can thus be associated with each other. This is a really important concept - if the nouns are related to each other, we can find that conceptual (lexical) chain in the content, even when those sentences are separated by many other unrelated sentences.
The “score” of a lexical chain is directly related to the length of the chain and the relationships between the chaining nouns (same word, antonym, synonym, meronym, hyper/holonyms, etc.)
Inside of Lexalytics software, theme extraction uses lexical chains for theme scoring. Summarization uses lexical chains to pick the most representative sentences. Entity sentiment scoring uses lexical chains to associate sentiment bearing phrases with the entities themselves.