Measuring the quality of industrial water used in a chemical plant using artificial intelligence (Logistic Regression, Naive Bayesian, Support Vector Machine, Random forest, and Decision Tree)

Articles in Press

Document Type : Original Article

Authors

1 Department of Chemical Engineering, Marvdasht Branch, Islamic Azad University, Marvdasht, Iran

2 Associate Professor of Chemical Engineering, Department of Chemical Engineering, Marvdasht Branch, Islamic Azad University, Marvasht, Iran

10.22034/ijche.2023.408953.1334

Abstract

Continuous measurement of factory water quality is important. Current methods of measuring water quality are not efficient enough. In this research, a new method using the concepts of artificial intelligence and machine learning has been proposed to solve the mentioned challenges. The proposed research method has been trained and validated using 472 samples of chemical data in MATLAB software. Each data sample has 6 input attributes (pH, conductivity, water hardness, total water-soluble solids, free chlorine, and alkalinity) and one output attribute (target). The parameters of disturbance matrix, precision, accuracy, and readability have been used to evaluate the efficiency of water quality measurement. The highest accuracy is related to the random forest method. The decision tree, simple Bayes, and vector machine methods are the same. The most refreshing rate is related to the decision tree method. The artificial intelligence method of the proposed decision tree with an accuracy equal to 70%, accuracy equal to 98%, and recall equal to 96% compared to logistic regression methods, Naive Bayesian method, support vector machine, and random forest, shows more efficiency and less error.

Keywords

Main Subjects

Iranian Chemical Engineering Journal

Articles in Press, Accepted Manuscript
Available Online from 10 October 2023

Files

Share

How to cite

Statistics