Using neural network in context awareness to improve email management

Wainaina, David Ngugi
Journal Title
Journal ISSN
Volume Title
This paper details a context awareness application which can be used in personal computers to enhance user experience. Users are distracted frequently and lose focus because there is so much information accessible through the computer. The users have to sift through for the information they really needed at that moment in time. Contexto application address this problem by analysing the information which comes to the system and decide whether it falls in the current user’s current context or not. It applies a neural network which uses sum of the weights assigned to each input to make the decision if a given threshold. The application addresses the problem by filtering emails received in outlook mail client. Each email which comes to the inbox will be analysed to determine which context it falls in and it is flagged for that context. The user can now read emails intended to the context he/she is working in efficiently without unnecessary disturbance. This is done by analysing the email’s sender and subject. The sender and words contained in the subject will have their own weights in a database. Each input’s weight is multiplied by the weight of input category; whether sender or subject keywords. The sum is then compared to a threshold and if it is more than the threshold, the algorithm will determine which context the email best falls in. The experiment demonstrated that it is possible to determine information context using neural network analysis and rearrange it to improve user experience. It is possible to extend the implementation for the email client to file system and browser to make a holistic user experience.
Applied project submitted to the Department of Computer Science, Ashesi University College, in partial fulfillment of Bachelor of Science degree in Computer Science, April 2016
user experience, email, Outlook, clustering