First, we conducted a detailed comparison of data mining tools
to select a suitable platform for conducting our research. We
began with a list of 30 data mining tools, which we filtered
down to 10 with good support for visual analysis. We then
applied the detailed methodology suggested by [2] to identify
a number of computational, functional, usability, and support
criteria necessary for this project. From the computational
perspective, we required that the system operate on a wide
range of platforms and be open source. This reduced the short
list further to three: Weka [1], Orange [3] and Yale [4].