Researchers develop new algorithm to detect misogynistic tweets

Researchers from Australia’s Queensland University of Technology have developed an algorithm to detect misogynistic tweets posted on the microblogging platform.

According to a paper that was published on springer.com, the team of researchers from the varsity collected data from Twitter’s streaming API that enables users to download data about tweets in real-time.

Through random sampling between January 2007 and March 2018 in the Australian domain, the team collected the tweets with “the three main misogynistic keywords (i.e. whore, slut, rape)”, the paper said.

The team finally obtained 5,000 tweets after removing those with many non-English words. Of these 5,000 tweets, the team had manually labelled 1,800 tweets as misogynistic. “We call this data as QUT Misogyny Identification (QMI) data.”

The researchers claim that misogynistic tweets detected by the algorithm will be 75 per cent accurate. With the help of the algorithm, the team says, social media firms can detect racist and homophobic contents too.

Delhi police to set up special cell to flag unlawful content

In a related new, as part of a pilot project, the Delhi Police is planning to set up an “IT cell” to flag “unlawful content” on social media that may have the potential to disrupt the law-and-order situation, officials said on Saturday (August 29).

They said the project will start once they receive an approval from the Ministry of Home Affairs (MHA) but for now, they have been allocated Rs 25 lakh for it.

“We are in the process of setting up an IT cell, which would start as a pilot project for three months and based on its outcome, we will take it forward, but for now, we are awaiting the approval of the MHA,” a senior police officer said.

The officer added that the aim is to flag unlawful content on social media, which may cause communal disharmony or affect the law-and-order situation, and take appropriate action against it.

(With PTI inputs)