StdDev refers to Standard Deviation
 Figure C.3: Classification errors for different RWIS sites obtained from predicting precipitation using classification algorithms.
Table C.4: Percentage of instances predicted correctly using classification algorithms (Experiment 4)
|
Actual Data
|
J48
|
RWIS Site
|
NP
|
P
|
NP->NPP
|
P->PP
|
P->NPP
|
NP->PP
|
19
|
86.43
|
13.57
|
81.99
|
5.59
|
4.44
|
7.98
|
27
|
76.29
|
23.71
|
68.78
|
13.31
|
7.51
|
10.40
|
67
|
63.65
|
36.35
|
50.74
|
18.55
|
12.91
|
17.79
|
25
|
84.92
|
15.08
|
80.49
|
6.61
|
4.43
|
8.47
|
56
|
86.09
|
13.91
|
81.31
|
4.73
|
4.78
|
9.18
|
60
|
79.59
|
20.41
|
71.95
|
8.82
|
7.63
|
11.59
|
68
|
87.74
|
12.26
|
85.31
|
5.29
|
2.43
|
6.97
|
78
|
83.33
|
16.67
|
78.30
|
9.46
|
5.03
|
7.21
|
14
|
78.48
|
21.52
|
73.66
|
10.33
|
4.82
|
11.19
|
20
|
70.49
|
29.51
|
60.79
|
15.77
|
9.70
|
13.74
|
35
|
84.15
|
15.85
|
78.15
|
6.22
|
6.00
|
9.63
|
49
|
86.41
|
13.59
|
82.47
|
5.90
|
3.94
|
7.69
|
62
|
85.87
|
14.13
|
82.03
|
6.99
|
3.84
|
7.13
|
Average
|
81.02
|
18.98
|
75.06
|
9.05
|
5.96
|
9.92
|
Std Dev
|
7.23
|
7.23
|
9.97
|
4.36
|
2.84
|
3.12
|
[table continued on the next page]
Std Dev: Standard Deviation
NP: No precipitation reported
P: Precipitation reported
NP->NPP: No precipitation predicted when No precipitation was reported
P->PP: Precipitation predicted when precipitation was reported
P->NPP: No precipitation predicted when precipitation was reported
NP->P: Precipitation predicted when precipitation was reported
Table C.4: Percentage of instances predicted correctly using classification algorithms (Experiment 4) [table continued from previous page]
|
Naive Bayes
|
Bayes Nets
|
RWIS Site
|
NP->NPP
|
P->PP
|
P->NPP
|
NP->PP
|
NP->NPP
|
P->PP
|
P->NPP
|
NP->PP
|
19
|
7.88
|
13.11
|
78.55
|
0.46
|
1.81
|
13.48
|
84.62
|
0.09
|
27
|
43.72
|
14.14
|
32.57
|
9.57
|
52.02
|
11.92
|
24.27
|
11.79
|
67
|
27.66
|
12.55
|
35.99
|
23.80
|
40.62
|
18.89
|
23.03
|
17.46
|
25
|
2.53
|
14.92
|
82.38
|
0.16
|
2.29
|
15.02
|
82.63
|
0.07
|
56
|
2.99
|
13.78
|
83.10
|
0.13
|
4.63
|
13.75
|
81.46
|
0.16
|
60
|
3.33
|
19.95
|
76.26
|
0.46
|
1.17
|
20.32
|
78.41
|
0.09
|
68
|
28.22
|
10.78
|
59.52
|
1.48
|
13.82
|
11.43
|
73.90
|
0.86
|
78
|
13.18
|
15.68
|
70.15
|
0.99
|
35.23
|
13.41
|
48.10
|
3.26
|
14
|
49.44
|
12.06
|
29.04
|
9.46
|
57.24
|
9.69
|
21.24
|
11.83
|
20
|
35.03
|
19.88
|
35.45
|
9.63
|
45.98
|
16.11
|
24.50
|
13.40
|
35
|
9.00
|
14.79
|
75.15
|
1.06
|
2.44
|
15.65
|
81.71
|
0.20
|
49
|
7.44
|
13.20
|
78.97
|
0.39
|
6.81
|
13.41
|
79.60
|
0.18
|
62
|
10.29
|
13.36
|
75.58
|
0.76
|
2.28
|
13.97
|
83.60
|
0.16
|
Average
|
18.55
|
14.47
|
62.47
|
4.51
|
20.56
|
14.39
|
60.46
|
4.60
|
Std Dev
|
16.31
|
2.73
|
21.19
|
6.98
|
22.01
|
2.90
|
27.49
|
6.47
|
Table C.5: Results obtained from using regression algorithms to predict visibility at an RWIS site (Experiment 5). Feature vector consists of temperature information from RWIS-AWIS sites in a set along with precipitation type and visibility for the RWIS sites. The table has mean absolute error values averaged over ten 10-fold cross-validations.
| |
ML Algorithms
| |
RWIS Site
|
LMS
|
LR
|
M5P
|
MLP
|
Set 1
|
67
|
1.6519
|
1.7436
|
1.6584
|
1.8020
|
Set 2
|
35
|
0.0785
|
0.1204
|
0.1024
|
0.1838
| |
49
|
0.0491
|
0.0756
|
0.0613
|
0.0643
| |
62
|
0.0511
|
0.0785
|
0.0618
|
0.0716
|
Set 3
|
56
|
0.0456
|
0.0722
|
0.0573
|
0.0975
| |
68
|
0.0160
|
0.0296
|
0.0244
|
0.0478
| |
78
|
0.1791
|
0.2081
|
0.0845
|
0.1545
|
 Figure C.4: Mean absolute errors for different RWIS sites obtained from predicting visibility using regression algorithms.
Table C.6: Percentage of instances with a certain distance between the actual and predicted temperature class value, obtained by using HMM to predict temperature class (Experiment 7).
|
Set 1
|
Set 2
| |
19
|
27
|
67
|
14
|
20
|
35
|
49
|
62
|
Distance 0
|
18.48%
|
16.18%
|
19.30%
|
21.94%
|
27.37%
|
22.86%
|
18.55%
|
22.37%
|
1
|
68.31%
|
28.67%
|
67.63%
|
73.11%
|
59.03%
|
65.06%
|
60.16%
|
70.88%
|
2
|
12.08%
|
49.85%
|
12.19%
|
4.83%
|
12.24%
|
11.09%
|
16.17%
|
4.38%
|
3
|
1.09%
|
4.59%
|
0.88%
|
0.11%
|
1.32%
|
0.95%
|
3.57%
|
1.15%
|
4
|
0.04%
|
0.71%
|
0.00%
|
0.00%
|
0.04%
|
0.04%
|
1.12%
|
0.53%
|
5
|
0.00%
|
0.01%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.35%
|
0.67%
|
6
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.08%
|
0.00%
|
7
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
8
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.00%
|
0.01%
| | | | | | | | | | |
Set 3
| | | | |
25
|
56
|
60
|
68
|
78
| | | |
Distance 0
|
42.65%
|
44.42%
|
39.60%
|
20.11%
|
39.69%
|
|
1
|
52.82%
|
49.00%
|
53.09%
|
27.20%
|
39.26%
|
2
|
4.33%
|
6.14%
|
6.96%
|
18.24%
|
16.86%
|
3
|
0.20%
|
0.42%
|
0.35%
|
8.36%
|
1.99%
|
4
|
0.01%
|
0.02%
|
0.00%
|
14.59%
|
0.51%
|
5
|
0.00%
|
0.00%
|
0.00%
|
7.17%
|
1.20%
|
6
|
0.00%
|
0.00%
|
0.00%
|
3.83%
|
0.49%
|
7
|
0.00%
|
0.00%
|
0.00%
|
0.46%
|
0.00%
|
8
|
0.00%
|
0.00%
|
0.00%
|
0.02%
|
0.00%
| |
Yüklə 1.32 Mb.