stats[statplots, scatterplot](data, format=excised[p], ..)
statplots[scatterplot](data, format=excised[p], ..)
scatterplot(data, format=excised[p], ..)
fraction of data points to remove from the plot
Important: The stats package has been deprecated. Use the superseding package Statistics instead.
The function scatterplot with the format parameter format=excised of the subpackage stats[statplots] removes the fraction of least dense data and plots a scatterplot of the remaining data.
This type of plot is often used when there is a large number of data points involved. The idea is that it is not necessary to have the detail of each individual point in a plot. Closely grouped points are supposedly more representative, so all other points are excised.
When p is a positive number between zero and one, p times the number of points are excised from the plot. In other words only the 1−p⁢numpoints most densely clustered points are plotted in the scatter plot.
When p is a negative number between zero and -1, p⁢numpoints of the most densely clustered points are excised from the plot. In this case 1−p⁢numpoints of the least densely clustered points are plotted in the scatter plot.
The default, when p is not specified is zero. This plot is identical to a scatter plot with no format parameter.
Relative density of points is calculated approximately by finding the distances between all points using the distance metric exp⁡−|p1−p2|.
Class data is converted to classmarks before generating the plot. Weighted data is accounted for. Missing data is ignored.
The command with(stats[statplots]) allows the use of the abbreviated form of this command.
data1 ≔ random[normald]⁡30,random[normald[3,1]]⁡20:
data2 ≔ random[normald]⁡30,random[normald[3,1]]⁡20:
data3 ≔ 12.00,Weight⁡10,3,8..9.5,9.67,11.11,10.34:
Download Help Document