An Investigation of the Evaluation and Predicting the Wastewater Quality Indicator Utilising Fuzzy and Artificial Neural Network Algorithms
Аннотация
The primary human activities responsible for surface water contamination are numerous. It is not always possible to determine the true condition of the water quality due to the unequal dispersion of industrial companies throughout the areas of Uzbekistan's major river basins. There are three goals for this article: First, the Uzbekistan technique for determining out the Water Quality Index was modified to account for the degree of toxicity of the most harmful chemical components. Secondly, fuzzy logic was used to model the Water Quality Index evaluation. Lastly, an artificial neural network model was developed to forecast the Water Quality Index. The fuzzy logic model used multiple input variables to produce a single output parameter (Water Quality Index). Seven artificial neural network algorithms with different their loss function and activation function optimizers were examined in the study's last phase. Applying Adam's loss function optimization and an artificial neural network (artificial neural network) with the softmax activation function, the best results were displayed (mean absolute percentage error 8.9%; R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> = 0.952). Satisfactory agreement was found by contrasting the created an artificial neural network algorithm's mean absolute percentage error and R2 metrics with those of earlier models, comparing the predicted and target information used to evaluate the quality of water. This study is interesting since it suggests changing the Water Quality Index evaluation approach that is already in use in Uzbekistan. Simultaneously, one may efficiently assess and forecast Water Quality Index values by using mathematical tools like the artificial neural network and the fuzzy logic approach in concert, sequentially and cooperatively.