Hierarchical Cluster with Maximum Plot Distance Help?

[epsilon80_1997]

Trying to figure out how to hierarchical cluster the maximum number of clusters of less than 3 units of difference. The data points can reach as much as a few thousand. Distance formula is (absolute value difference of x values + one-half of absolute value of y values, unless the y difference is greater than the absolute value of the x value difference, then use the full absolute value of the difference of the y values) (for example abs(x-x) + (abs(y-y))/2) = distance, but if abs(y-y) is larger than (x-x), then abs(x-x) + (abs(y-y). Example distance results as well:

Distance from 16384, 16384 x y Distance 16384 16384 0 16385 16385 2 16386 16386 3 16384 16387 3 16383 16381 3 16381 16384 3 16387 16385 4 16381 16386 4 16384 16379 5 16381 16379 5 16383 16389 5 16386 16389 5 16381 16381 5 16380 16379 5 16380 16382 5 16388 16378 6 16388 16387 6 16385 16390 6 <colgroup><col><col><col></colgroup><tbody> </tbody> <colgroup><col><col><col></colgroup><tbody> </tbody>

<colgroup><col span="2"></colgroup><tbody>
</tbody>

 

[epsilon80_1997]

		16384	16385
		16384	16385
x	y	distance	distance
16390	16386	7	6

<colgroup><col width="64" span="4" style="width:48pt"> </colgroup><tbody>
</tbody>

an example of the distance results


formulas used:

		16384	16385
		16384	16385
x	y	fields
16390	16386	=ROUNDUP(IF(ABS($B272-D$2)/2>ABS($A272-D$1),ABS($B272-D$2),ABS($B272-D$2)/2+ABS($A272-D$1)),0)	=ROUNDUP(IF(ABS($B272-E$2)/2>ABS($A272-E$1),ABS($B272-E$2),ABS($B272-E$2)/2+ABS($A272-E$1)),0)

<colgroup><col span="2"><col><col></colgroup><tbody>
</tbody>