Similarity Relations
<img src="similarity.jpg" width="350" height="150" border="0"> For more information:

Forte

Morris

Teitelbaum

Isaacson

Broline

Instructions

Click inside the browser window to direct keyboard input to the program.

Using the number keys or numeric keypad, enter vector data into the first vector field. The left and right arrow keys and the delete key can be used for editing. The vector must be completed before proceeding.

To select the second vector field, click it, or press the tab key. Enter the data for the second vector.

When both vector fields are filled, press the return or enter key, or click the name of one of inventors of the various similarity functions to complete the analysis.

Click a field or press the tab key to edit or replace the data in either field, then continue as above.

Forte: “R”
R0: minimally similar (no corresponding entries)
example: 221100, 112011
R1: maximally similar with interchange (4 identical entries; remaining switched)
example: 221100, 212100
R2: maximally similar without interchange (4 identical entries)
example: 223111, 222121
top
Morris: “SIM”
the sum of the absolute values of the differences between corresponding entries in two interval vectors

Example
vector 1: 030201
vector 2: 121110

|0 - 1| + |3 - 2| + |0 - 1| + |2 - 1| + |0 - 1| + |1 - 0|

= |-1| + |1| + |-1| + |1| + |-1| + |1|

= 1 + 1 + 1 + 1 + 1 + 1

= 6

top
Teitelbaum: “similarity index”
the square root of the sum of the squares of the differences between corresponding entries in two interval vectors

Example
vector 1: 030201
vector 2: 121110
=
=
=
= 2.45
top
Isaacson: “IcVSIM”
the population (n-weighted) standard deviation of the differences between corresponding entries in two interval vectors

Example
vector 1: 543210
vector 2: 210000

step 1
find the “difference vector,” a six-element array containing the differences between the corresponding interval vector entries:
step 2
if you have a scientific calculator, find the standard deviation of the six elements in the difference vector (here’s how I do it on my TI-36X):
but it’s more involved if you can’t do the statistics on your calculator:

step 2a
find the mean of the difference vector:
step 2b
find the difference between each difference vector entry and the mean, square it, and sum all the results:
step 2c
take the square root of the previous sum divided by 6:
top
“Broline’s scaled taxi-cab metric”

Example 1

vector 1: 122230 (10 intervals)
vector 2: 114112 (10 intervals)

Since the vectors contain the same total number of intervals, we only have to find the sum of the absolute values of the differences:

|1-1| + |2-1| + |2-4| + |2-1| + |3-1| + |0-2| = 0 + 1 + 2 + 1 + 2 + 2 = 8

and then divide it by 10 to scale the result:

8/10 = .80



Example 2

vector 1: 122230 (10 intervals)
vector 2: 030201 (6 intervals)

Since the vectors contain different numbers of intervals, we scale each vector and then find the sum of the absolute values of the differences:
top