A new computational linguistics tool from OpenAmplify is promising to make life simpler by analysing large amounts of data to find out users' real wants and needs.
The TopicIntentions application analyses online conversations, web content and forum discussions and uses linguistics programming to find out what is important to particular users.
For example, it can help advertisers identify the meaning and tone of web media before they post adverts around it.
Michael Petit, co-founder and chief information officer of OpenAmplify, told V3.co.uk: “This will make a big splash in digital media and make for more intelligent targeting that makes advertising more relevant.
“Brand safety is ensured. You won’t see those horrible but amusing examples of airplane adverts appearing next to a story about an air crash, we can make that a thing of the past.”
Another use could be to moderate forum comments automatically, since the software could be set to look for worrying language by both looking for key words but also the context in which they are used for more accuracy.
Search engines too could benefit from the tool, according to Petit. He gave a personal example – he is currently building a boat but a search for boat repair terms brings up a majority of returns consisting of people asking similar questions, rather than answers.
Social media is another area that Petit expects to be most interested in the tool.
“Look at social media, there’s a massive goldmine of content out there about people’s views and opinions," said Petit. "The problem is how you mine that stuff.”
Antarctica lost on average 252 gigatons of ice mass per year from 2009 to 2017, claims study
Buyers can demand refunds if they've had a game for no more than 14 days and not registered more than two hours of play
Total lunar eclipse 2019: 'Super Blood Wolf Moon' to be visible across Europe and North America on Sunday night
Moon will turn reddish-orange in colour during this weekend's total lunar eclipse
Hackers to compete for prize money of between $35,000 and $250,000 cracking the Tesla Model 3 at this year's Pwn2Own contest