Personality-aware interfaces for learning applications

  • Authors:
  • Nicholas Caporusso

  • Affiliations:
  • IMT - Institute for Advanced Studies, Lucca, Italy

  • Venue:
  • Proceedings of the 37th annual ACM SIGUCCS fall conference: communication and collaboration
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Although several personalization features are included in the design of e-learning applications, both the personality and the cognitive requirements of the users are among the last characteristics that are considered. As a result, almost all software usually present contents in exactly the same fashion to all users, without taking into account their different attitudes. Conversely, cognitive theories declare that the learning process can be improved by tailoring the way in which information is conveyed. Unfortunately, many parameters have to be considered to obtain reliable results, and processing personality is computationally expensive. In this study, we introduce a light-weight framework for eliciting personality traits and translating them into user interface characteristics which adapt on-line to the diverse learning styles. Our system recognizes the heterogeneity of cognitive requirements, and simultaneously segments user populations accordingly. Also, a personality-aware interface automatically reconfigures its information presentation layer in real-time, accommodating the character traits acquired from the users, to improve the learning performance. We evaluated the bi-directional relationship between personality and interface in the context of web-based e-learning software for customer management services. Our experimental study focused on the mutual influence between the organization of the information presentation layer and the cognitive attitude of the users. Results show that, by choosing the appropriate parameters and by tuning the interface accordingly, the learning curve can be reshaped so that the time required to learn both the application features and the content can be shortened by approximately 30%.