in order to be accurate, it would have to gauge the intention behind the posts. therefore it would need to gauge intentional bashing too. because of that, bashing could be seen as a sign of desperation. the bashers might have the opposite effect they intend to have.
i believe this is called 'machine learning'. it uses unstructured data to learn from. the 'watson' program that beat the human champions at 'jepoardy' used this. it has evolved into other programs. one of which is called 'watson discovery'. it sounds like something close to what 'stocksonar' uses.
we are about to start learning things faster than humans can keep up with. much faster than politicians can keep up with. there has already been a discovery in science that humans haven't been able to make, that machine learning has made. those discoveries are about to mount up and they are just barely getting started.
but in the end, as i have said right here on this forum, sentiment is what always drives stocks until or unless they are bought out. with that in mind, what ever you write about a stock does always have an impact even if it is only slight.
