Critique of User-Adaptive Information Visualization

Hi Abed,
Thank you very much for reading my page and your kind suggestions:

  1. I just added a few lines, explaining the source of their decision. However, if you mean, why they chose 5 and not 6 for instance, there was not any explicit reason presented in the paper, but I assume they simply wanted to keep the number of variables managable, and the reason they chose that particular set was because of their higher significance in the study of Pekrun which I referred to.
  2. Machine learning (ML) models (classifiers) are the core elements of adaptive systems, and ML is categorized under AI. For example, in one study, they analyzed eye-gaze data to create user- and task model; they used a classifier to classify different affective states based on the eye-gaze features related to the area of interests in a visualization. In every U-AIV system, there is a user or task modeling phase, which provides the basis upon which they can decide when to present the adaptive visualizations or adaptive support in general. As you mentioned presenting those feedbacks should be based upon a user model or task model and that's where we can see the most AI techniques and ML models are used. The other AI’s direct techniques and artifacts such as sequence planning models (HMM, Max Entropy Markov Model, etc.), all come very handy when we discuss the advanced type of problems in User Adaptive Information Visualization. However, as the purpose of this page was to introduce the field by two papers from two different research teams, I decided to select these two papers, because they present a nice introduction to the type of problems that are investigated in U-AIV, and without this foundation, discussing the advanced problems would be a bit pointless.
  3. I just added a dedicated section (under introduction), in which, I elaborated this relationship.


Thanks again for your valuable remarks, I really appreciate it ;)
Yaashaar

Yaashaar HadadianPour (talk)10:07, 14 March 2016