Databases Reference
In-Depth Information
Figure 2.22
PCP Pollen Dataset
Both attributes appear to be somewhat normally distributed, yet both have
some irregularly shaped (not smooth) peaks. When rotating to the right or left it
is easy to see the distributions of the attributes on either end, but not so easy to
see distributions of attributes in the middle. To clearly see the distributions of
attributes currently displayed on the middle axes you would need to move them
to the outside.
To do this, hover over the axis of the attribute that you wish to reposition,
then drag it to the right or left. You can tell when you are hovered over
an axis when you see the attribute name enlarge and change from white
to yellow.
Move the ObsNo attribute all the way to the right.
Look at its distribution. What do you see? The distribution is flat because
observation numbers are a sequential numbering starting with 1 and going to
3,848. Obviously, this attribute should not be included in any data mining
operation, since it should have no relationship with any of the other attributes.
Eliminate ObsNo from the plot by again hovering over the axis and right-
clicking. Select “Hide ObsNo” from the context menu.
You should now see a plot similar to Figure 2.23. Notice the difference in
pattern between line segments connecting Weight and Ridge versus the patterns
