Saif Mohammad and his co-workers from the National Research Council Canada, uncover feelings inside an email by using sentiment analysis.
A huge "sentiment" database was created by the team by using crowdsourcing by hiring employees for Amazon's Mechanical Turk to pair 24,200 words with emotions for instance "ice cream" was paired with "joy."
Comparing the words used in 32,045 emails that were made public after the Enron scandal, helped make it possible to assess emotional tone and how it varies with gender, New Scientist reported.
The findings of the study showed a marked difference between sexes.
Women mostly used words that were related to joy or sadness, while men favoured words that were associated with fear or trust.
However, both sexes used joyous and cheerful words when they wrote to women.
The team also tried to find out personality traits from emotional content.
--ANI (Posted on 28-09-2013)