Towards Composable Prediction of Contact Groups

This research project focuses on the creation and testing of pre-generated groups of contacts that contain e-mail addresses that the user is likely to send e-mail to as a group in the future. The purpose of these generated groups is to make it as easy as possible for the user to send a message to a group of people without modifying the group as much as possible. Some groups may be very large and thus will make sending e-mail more efficient as they will save the user from a lot of typing. These groups are automatically generated and recommended to the user, who will accept the group as it is, modify it then accept it, or reject it if it does not mean anything to the user. After coming up with several different algorithms to generate the groups, we tested them by running a user study and collecting data from 19 participants. We then calculated the average percentage difference between a predicted group and a future message. This allows us to see how viable our groups are and how likely they are to help a user compose an e-mail message in the future. We concluded that using Google score was the best method.Bachelor of Scienc