Skip to Main Content

Analysis of variance (ANOVA) can be used to further examine the model's feasibility and validity (Tables 4 & 5). F-value (Fisher variation ratio), lack of fit, p-value (probability value), adequate precision (AP), R2d (coefficient of determination), R2Adj (adjusted coefficient of determination), R2Pred, were some of the evidences. AP corresponds to signal to noise ratio, which predicts contrasts, the range of expected values at nodes to the prediction error. Model selectivity is adequate when the ratios are higher than four (Zhou et al. 2011; Mishra et al. 2021; Sawood et al. 2021b).

Table 5

ANOVA for response surface quadratic model

SourceSum of squaresdfMean squareF-valuep-value
Model 8.153E + 005 90,587.12 270.24 <0.0001 significant 
A-Influent As(V) concentration 15,612.35 1 15,612.35 46.58 <0.0001  
B-Inlet flow rate 8,273.95 1 8,273.95 24.68 0.0006  
C-Bed height 7.484E+005 1 7.484E+005 2,232.51 <0.0001  
A2 27.66 1 27.66 0.083 0.7798  
B2 983.96 1 983.96 2.94 0.1174  
C2 18,866.96 1 18,866.96 56.28 <0.0001  
AB 13.78 1 13.78 0.041 0.8434  
AC 4,816.71 1 4,816.71 14.37 0.0035  
BC 117.81 1 117.81 0.35 0.5665  
Residual 3,352.06 10 335.21    
Lack of fit 3,352.06 4 838.01    
Pure error 0.000 6 0.000    
Cor total 8.186E + 005 19     
SourceSum of squaresdfMean squareF-valuep-value
Model 8.153E + 005 90,587.12 270.24 <0.0001 significant 
A-Influent As(V) concentration 15,612.35 1 15,612.35 46.58 <0.0001  
B-Inlet flow rate 8,273.95 1 8,273.95 24.68 0.0006  
C-Bed height 7.484E+005 1 7.484E+005 2,232.51 <0.0001  
A2 27.66 1 27.66 0.083 0.7798  
B2 983.96 1 983.96 2.94 0.1174  
C2 18,866.96 1 18,866.96 56.28 <0.0001  
AB 13.78 1 13.78 0.041 0.8434  
AC 4,816.71 1 4,816.71 14.37 0.0035  
BC 117.81 1 117.81 0.35 0.5665  
Residual 3,352.06 10 335.21    
Lack of fit 3,352.06 4 838.01    
Pure error 0.000 6 0.000    
Cor total 8.186E + 005 19     

Close Modal

or Create an Account

Close Modal
Close Modal