Would be interesting to run this algorithm on published scientific papers (another group recently claimed a different algorithm as being able to detect schizophrenia purely from text) and speech from politicians.
Yes, but to be fair, the article does go on to say:
"...their work poses several outstanding questions. For one thing, their sample size of 34 patients was tiny. Researchers are planning to attempt to replicate their findings using transcripts from a larger cohort of at-risk youths."
Comments
Would be interesting to run this algorithm on published scientific papers (another group recently claimed a different algorithm as being able to detect schizophrenia purely from text) and speech from politicians.
Interesting concept, but 'analysis is flawless' with '100% accuracy' on a sample size of 30-something? C'mon.
Yes, but to be fair, the article does go on to say:
"...their work poses several outstanding questions. For one thing, their sample size of 34 patients was tiny. Researchers are planning to attempt to replicate their findings using transcripts from a larger cohort of at-risk youths."