Mental sentiments are spotted by a fresh computerised learning system, like satire and sarcasm, in e-mails and text messages and it may even find content that indicates suicidal ideations.
After taking a class in artificial intelligence Saig developed the system in the Technion’s Learning and Reasoning Lab.
Based on Saig, inflections and voice tone play an essential job in expressing one’s significance in verbally conveyed message.
In text and e-mail messages, those nuances are lost and writers who wish to signify sarcasm, empathy or uncertainty have taken to using pictures, or “emoticons,” like the smiley face, to compensate.
“These icons are superficial clues at best.
Lately, pages meant to be humourous on social networks like Facebook and Twitter were titled “first-class and condescending folks,” or “normal and reasonable individuals.” Such pages are popular in Israel, said Saig, and users are encouraged to submit ideas for phrases that may be tagged as’ expressions that were stereotypical,’ for that specific page.
Saig identified present routines by finding posts to these groups. The technique he developed enables the system to find routines that are future on any social network.
Saig implemented ‘Machine Learning’ algorithms to the content on those pages and used the results to automatically identify stereotypical behaviors discovered in social network communicating.
The system was built to identify key words and grammatical customs that were feature of sentence structure suggested by the sentiments of the content.
“Now, the system can recognise patterns which are either condescending or caring thoughts and may also send a text message to the user in the event the system believes the post could be pompous,” said Saig.
When applied to other networking pages it might help find content that indicates suicidal ideations, as an example, or ‘calls’ for help, or expressions of happiness or admiration, Saig said.