Table 2 The parameter selection results of the WQImin models from the stepwise multiple linear regression based on the training dataset (n = 264).

From: Spatiotemporal variation evaluation of water quality in middle and lower Han River, China

Parameter selection

WQImin-w (weighted)

WQImin-nw (non-weighted)

Models

R2

PE(%)

P

Models

R2

PE(%)

P

Zn

w1

0.408

9.909

 < 0.001

nw1

0.408

9.909

 < 0.001

Zn, PI

w2

0.534

5.595

 < 0.001

nw2

0.573

4.395

 < 0.001

Zn, PI, NH3N

w3

0.469

6.531

 < 0.001

nw3

0.558

5.278

 < 0.001

Zn, PI, NH3N, TP

w4

0.597

10.759

 < 0.001

nw4

0.664

8.472

 < 0.001

Zn, PI, NH3N, TP, DO

w5

0.780

6.699

 < 0.001

nw5

0.815

5.681

 < 0.001

Zn, PI, NH3N, TP, DO, Pb

w6

0.827

3.725

 < 0.001

nw6

0.849

3.439

 < 0.001

Zn, PI, NH3N, TP, DO, Pb, Cu

w7

0.915

3.146

 < 0.001

nw7

0.926

2.642

 < 0.001

Zn, PI, NH3N, TP, DO, Pb, Cu, COD

w8

0.955

3.820

 < 0.001

nw8

0.958

3.302

 < 0.001

Back to article page