Aldo Faisal and Gabriela Tavares of Imperial College London studied more than 160,000 tweets from personal accounts held by individuals, "managed" accounts belonging to large, well-known corporations and "bot-controlled" accounts.
Writing in the journal PLoS ONE, the researchers say they found periods of high or low Twitter activity and the time between successive tweets could identify the three kinds of accounts with approximately 83 percent accuracy.
Corporate-managed accounts tweeted more during work hours, they said, personal accounts were more active in the afternoons and evenings, while bot-controlled accounts either tweeted at regular, constant intervals throughout the day or had sudden bursts of activity at a few specific hours.
"The identification and classification of specific types of users on Twitter can be useful for a variety of purposes, from the computational social sciences, focusing advertisement and political campaigns, to filtering spam, identity theft and malicious accounts," senior study author Faisal said.
Duggar sisters unveil Christian dating rules in new book
LGBT community has 'bullied the American people': Bachmann