|
Performance evaluation of adaptive neuro-fuzzy inference system for modelling dissolved oxygen of Kubanni Reservoir: A case study in Zaria, Nigeria
|
Eneogwe Chukwuemeka    , Sanni Ismaila Mohammed , Abubakar Alfa Umar , Idoko Apeh Abraham , Bello Abdulrazaq Ayobami |
| Corresponding author: Department of Water Resources and Environmental Engineering, Ahmadu Bello University, Zaria, Nigeria , emekaeneogwe@gmail.com |
|
|
Abstract: (1170 Views) |
Background: Water quality evaluation require arduous laboratory and statistical analyses comprising of sample collection and sometimes transportation to laboratories, which may be expensive. In recent years, there has been an emergent need to monitor the dissolved oxygen (DO) concentrations of Kubanni reservoir as a result of anthropogenic and agricultural pollution. Hence, this study was conducted to apply adaptive neuro-fuzzy inference system (ANFIS)-based modelling in the prediction of DO of Kubanni reservoir.
Methods: Water quality data for seven years were used to develop ANFIS models. Six water quality parameters, namely, total dissolved solids, free carbon dioxide, turbidity, temperature, manganese, and electrical conductivity, were selected for analysis based on their sensitivity. Subtractive clustering and grid partitioning techniques were considered when generating the fuzzy inference system (FIS). Three ANFIS models according to different lengths for training data and testing data were selected for modelling.
Results: The results showed that Model-1 gave the best correlation (R-squared and adjusted R-squared of 0.852503 and 0.845000, respectively) for whole data using six input variables. While Model-3 gave the best correlation (R-squared and adjusted R-squared of 0.807791 and 0.799940, respectively) for whole data using three input variables.
Conclusion: The performance efficiency of ANFIS model 1 using 6 inputs shows that the model is reliable for modelling water quality. |
|
| Keywords: Dissolved oxygen, Water quality modelling, Manganese, Nigeria |
|
|
Full-Text [PDF 811 kb]
(749 Downloads)
|
Type of Study: Original Article |
Subject:
Special
Received: 2022/10/15 | Accepted: 2022/10/15 | Published: 2022/12/22
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
