Page 30 - Special Topic Session (STS) - Volume 1

STS346 A.H.M. Rahmatullah Imon
at the upper or lower corner of the plot depending on their signs; large
positive outliers will be plotted at the upper corner and large negative outliers
will be located at the bottom corner of the plot.

3. Results
In this section we would like to present an example to demonstrate how
our proposed method works in the classification of spatial regression outliers
in both x and y dimensions. Here we consider a spatial outlier data given by
Hadi and Imon (2018) extending the idea of Shekhar et al. (2002). We present
the data in Table 1 and also in Figure 2.

Table 1: Hadi and Imon (2018) spatial outlier data
Index Location Attribute Diff_Location Diff_Attribute
1 1.0 2.0 * *
2 2.0 3.0 1.0 1.0
3 2.1 3.2 0.1 0.2
4 2.6 7.0 C 0.5 3.8 C
5 3.0 4.0 0.4 -3.0 C
6 3.8 5.0 0.8 1.0
7 3.9 5.6 0.1 0.6
8 4.0 5.7 0.1 0.1
9 4.2 1.6 D 0.2 -4.1 D
10 4.5 6.0 0.3 4.4 D
11 5.0 6.2 0.5 0.2
12 6.0 8.0 A 1.0 1.8
13 6.2 6.3 0.2 -1.7
14 6.4 6.1 0.2 -0.2
15 6.7 5.5 0.3 -0.6
16 7.1 5.0 0.4 -0.5
17 7.3 4.4 0.2 -0.6
18 7.5 4.3 0.2 -0.1
19 7.7 6.9 E 0.2 2.6 E
20 8.0 2.8 0.3 -4.1 E
21 8.4 2.1 0.4 -0.7
22 9.0 1.0 B 0.6 -1.1
23 9.2 2.1 0.2 1.1
24 10.0 2.7 0.8 0.6
25 10.1 3.2 0.1 0.5
26 11.0 4.0 0.9 0.8
27 15.0 F 4.1 4.0 F 0.1
28 17.0 4.2 2.0 0.1
29 19.0 4.3 2.0 0.1
30 20.0 4.4 1.0 0.1

19 | I S I W S C 2 0 1 9