Adaptable & adaptive personalised news recommendation interface

Abstract

“There is no such thing as an average user”. Efficiency and satisfaction in the user interface comes from an understanding of who the target users are, and thus knowing their overall context of the system use. On the other hand, it is very likely that users will differ in the content they require whilst interacting with a program. This thesis experiments with user-centred design, leading to the creation of an adaptable and adaptive interface, aimed to increase efficiency and user satisfaction, without impairing overall usability caused by complexity. In this project, we make use of ‘Monarch’, a newsreader in the form of a chrome extension, created in a previous FYP by Alexei Debono. Qualitative and quantitative data, acquired through a field study and a survey on potential users, are studied to shed light on the main segments of users contained within. Reliable and realistic representations of these segments are created in the form of Personas. With the personas created, potential interfaces are inferred and implemented within one interface. This interface is made adaptable, giving a user the ability to customize elements within the interface and thus forming a personalised interface. This implemented adaptable interface served as the starting point for the second part of the project which consisted of implementing adaptive functions in the interface. These functions aim to aid the user whilst editing the interface by automatically adjusting elements within the interface. With the use of usability testing and web usage mining, usage patterns are found and formed into association rules. This process is used as a basis for two adaptive functions implemented within the final interface, one providing adaptations based on usage data gathered in usability tests and the other based on usage data learnt from the current user. Thus, an adaptable and adaptive interface is created. Through data gathered in usability testing, we statistically prove that the adaptive functionality significantly improves efficiency with respect to time, and user satisfaction, without significantly affecting the overall usability of the system.