nlp - when calculating the cooccurance of two words, do we sepate the sentences or linking all sentences?

Question

For example, I get I document that contains 2 sentences: I am a person. He also likes apples. Do we need to count the cooccurrence of "person" and "He" ?

Answer 1

每个文档用换行符分隔。共现的上下文窗口仅限于每个文档。

基于这里的实现。

换行符被视为指示新文档（上下文不会跨越换行符）。

所以，根据你准备句子的方式，你可能会得到不同的结果：

设置 1：('He', 'person')同时发生

...
I am a person. He also likes apples.
...

设置 2：('He', 'person')未同时发生

...
I am a person. 
He also likes apples.
...