PWARX Model Identification Based on Clustering Approach - Complex System Modelling and Control Through Intelligent Soft Computations - page 175

(a)

1.5

(b)

1.5

estimated y
real y

real output
estimated output

1 1

0.5 0.5

0 0

−0.5 −0.5

−1 −1

0 50 100 150 200 250

0 50 100 150 200 250

k k

(c)

1.5

real output
estimated output

1

0.5

0

−0.5

−1

0 50 100 150 200 250

k

Fig. 2 The estimated outputs a with Chiu, b with DBSCAN, and c with k-means

Table 2 Validation results

Quality measures Chiu DBSCAN k-means

D h 0.0876 0.1514 0.1828

2
e

0.0023 0.0097 0.0109

r

FIT 89.4191 79.0183 74.195

For the DBSCAN based method, the choice of the synthesis parameters n q ,
MinPts and

e

is as follows:

<

n q ¼ 20 ;

MinPts ¼ 35 ;

:

e ¼ 0 : 85