AMA Assignment Multidimensional Scaling Avijit Banerjee Sec B, 170103048
Dataset – HBAT_MDS Case Processing Summary Cases
180
Sources
18
Objects
10
Proximities
Total Proximities
810a
Missing Proximities
0
Active Proximitiesb
810
a. Sum over sources of all strictly lower-triangular proximities. b. Active proximities include all non-missing proximities.
Interpretation: a) The no. of objects to be considered are 10 b) With respect to the 10 objects, data is read from 18 sources c) The total no. of cases considered for the scaling is 180
Goodness of Fit
Stress and Fit Measures Normalized Raw Stress
.08328
Stress-I
.28858a
Stress-II
.81673a
S-Stress
.21181b
Dispersion Accounted For (D.A.F.) Tucker's Coefficient of Congruence
.91672
.95746
PROXSCAL minimizes Normalized Raw Stress. a. Optimal scaling factor = 1.091. b. Optimal scaling factor = .956.
Interpretation a) The value of stress is 0.083 and is acceptable b) Stress is minimized when the objects are placed in a configuration such that the distances between the objects best match the original distances c) A trade-off must be made between fit of the solution and the number of dimensions d) The stress value varies for different dimensions; hence we have taken the number of dimensions to be 2 for an acceptable stress value
Common Space
Final Coordinates Dimension 1
2
var1
-.488
-.246
var2
-.467
-.435
var3
-.009
-.504
var4
.739
.120
var5
.650
-.285
var6
.222
.405
var7
-.534
.456
var8
.173
.818
var9
.325
-.449
var10
-.611
.121
Interpretation: a) The above table represents the coordinates for the various objects taken in the twodimensional space b) Here the two dimensions form the axes of the plot c) The measures above have been calculated on the basis of Identity scaling model
Interpretation: a) All the objects are plotted and represented graphically here b) The objects closer to each other have the minimum dissimilarity, whereas the ones far apart have maximum dissimilarity