Overcoming the email avalanche
Overwhelmed with emails? A JCU researcher wants to make it easier to automatically categorise emails based on their contents.
Dr Sisi Liu’s PhD looked at ways to prioritise which emails are more important, using an algorithm to analyse the sentiment behind the email.
“It can be used to prioritise, sort tasks, and communicate routine tasks,” she said. “Or for businesses dealing with customers, it can categorise and identify which emails to pay more attention to, such as prioritising negative feedback that needs addressing.”
Dr Liu said email is very much a part of modern life and she wanted to develop a way of making it more manageable.
“There’s a massive amount of communication that happens via text, and I recognised a need for making that easier in a business context,” she said.
“Despite the fact that email is a widely adopted contemporary means of communication in business settings, email sentiment hasn’t been thoroughly studied.”
Dr Liu’s main challenge was to improve the accuracy of algorithms that predict the emotions underlying emails.
“I was able to improve the accuracy of email sentiment detection to between 70 and 80 per cent,” she said. “Previously it was as low as 50 per cent, which is not very reliable.”
Dr Liu said it could also be used to track workplace morale.
“There’s an increasing focus on wellbeing in the workplace,” she said. “This technology could be used to track emotional changes in the office to get a company-wide perspective of how employees are feeling, and alert management when there are problems that need fixing.”
Dr Liu is currently working on prototype software for commercial application.
Dr Sisi Liu