Chapter 11 . Input Selection for an Algal Bloom Model
225
Table 11.1. Input Subsets Identified Using the Unsupervised Input Determination
Techniques
No.
apriori
of
identification
inpu
ts
Anabaena lags 1,2, 4
3,4
Turb. lags 1,2, ... ,
10
10
Col. lags 1, 2, ... , 10 10
Temp. lags 1,2,3,4 4
Flow lags 1,2, ... ,
10
10
pH lags 1,2
2
Total. P. lags 1,2,
10
... ,10
Sol. P. lags 1,2, ... , 10
10
TKN lags 1,2, ... ,
10
10
Total No. of InEuts
70
PC, principal component.
u prio/"l
identification
PCA
Anabaena PCs 1,2,
... ,7
Turb. PCs 1,2,3,4
Col. PCs 1,2,3
Temp. PCs 1, 2
Flow PCs 1,2,3
pH PCs 1,2,3,4
Silica PCs 1,2,3,4
Total. P. PCs 1,2,3,
4,5
Sol. P. PCs 1,2, ... ,6
TKN PCs 1,2, ... ,7
River Lev. PCs 1,2,3
Data
Dh Ided Using
P A
No.
No.
of
SOM
of
inpu
inpu
ts
ts
Anabaena lags 1,5,6,
7
9, 12, 13, 18, 20, 23,
11
24,25
4
Turb. lags 1, 14,20
3
3
Col. lags 1,20
2
2
Temp. lags 1, 13
2
Flow lags 1,3,5,7,8,
3
10, 11, 14, 15, 17, 18, 17
20,21,22,23,24,25
4
pH lag 1
4
Sol. P. lag 1
5
TKN lag 1
6
River Lev. Lag 1
7
3
48
39
QM
Figure 11.1.
The input determination techniques utilised to determine the
inputs for each of the six models.