NLP Entity Injection and Salience
Measuring which entities a page is actually about, by prominence rather than frequency, and correcting the page when the answer is not the one you intended.
Analogy Mental model
A film titled after one character, in which a supporting role occupies eighty percent of the screen time, leaves the audience unsure who the protagonist was.
In practice What you actually do
Paste a finished article into a natural language analysis interface and read the ranked entity list with its salience scores. If the top entity is an incidental example rather than your subject, move the subject into grammatical subject position and repeat it in the opening and in section headings, then measure again.
Where people get burned Pitfalls
This is a far better instrument than keyword density, because it asks whether the term occupies the subject or object position in the sentence, not merely how many times it appears.
This entry was written in English from the Chinese original, not machine-translated. Read the Chinese version →