Why is it unreliable to say that "judging whether Cao Xueqin wrote the forty chapters after a dream of red mansions by machine learning"?

This method is based on statistical methods, depending on the author's habit of choosing words and making sentences, and whether it was written by one person.

In fact, this method is feasible to some extent. But when this person did this research, the choice and number of word samples were controversial.

At the same time, I personally don't think it is necessary. Because if you read a dream of red mansions carefully, there are not only differences in words, but also differences in thoughts. There is no need to use this means.

In the first eighty times, even the girls in Jia's family ate exquisite snacks and didn't eat big fish and meat at all. The last forty times were fish and meat. In the first eighty times, Baoyu was close to his body, and everything in his study was from the west. Forty times later, the study was full of Wan Wen antiques. I never wrote the color of Daiyu's clothes in the first eighty times. The characters are romantic, so why wear colors? The last forty times, red and green came.

Different worldviews are really different, and so are seeing and not seeing.