RésuméNowadays, data sets are not only bigger in their size but also in their dimensionality. Interactive visualization of highly multi-dimensional data sets is crucial to present and discover the insight, pattern and (ir)regularities of information. The classical methods are not usually efficient to handle medical data sets with so many attributes. This thesis presents a tool to visualize and explore a data set that contains several thousands of attributes. This tool works in two phases. The first one is the reduction of the amount of attributes, through the construction of data groups, transformed in turn into interval data. The second phase visualize the low multi-dimensional data using 3-dimensional multiple scatter plot. The system also provides interactive exploration so that analysts can view, interact and manipulate the processed data
as well as the information in its original form.
|la date de réponse||2010|
|Superviseur||Monique FRAITURE (Promoteur)|