Friday, 29 July 2016

On 01:51 by admin   No comments
Cyberbullying and post -an offensive in cyberspace is an unfortunate reality. Scientists Yahoo Twitter seems to support measures to protect users by making internet abuse detection equipment.

As reported by Ubergizmo , Friday (07/29/2016), Yahoo Labs has created an algorithm that automatically filter out abuse. Most filters work by detecting the words, phrases and expressions commonly used to mem- bully someone.
They also apply machine learning to record coarse or offensive comments marked by Yahoo editors. This allows the algorithm to process the words as 'vectors' instead of treating words as good or bad. As a result, the algorithm can choose a string of words that are considered offensive, even if the comments are not using words deemed offensive.
So far, the team has had a pretty good success with their algorithm and has had a 90 percent accuracy rating. It is not yet clear whether Yahoo will implement this system or not.

For your information, a British researcher Alex Krasodomski previously mentioned that identify the offending nudge and not actually very difficult. "If given 10 tweet , a group of humans rarely all would agree on a tweet should be classified in abusive, so you can imagine how difficult it is to do the computer," he added.

0 comments:

Post a comment