Water Quality Prediction Using Random Forest Algorithm

Abstract

Water is an essential requirement for every living thing in the globe. The massive growth of technology helps to attain water purity and quality for the betterment of people. Because of industries, chemical wastes mingle with water resources by contaminating water and land. People who drink the contaminated water suffer from waterborne diseases like cholera, Diarrhea, typhoid, etc. Water scarcity has been rapidly increasing with the increase in population and pollution. Nearly two lakh people are dying in India due to water scarcity and waterborne diseases. So we came forward to solve a complex problem regarding water. This article suggests a machine learning-based method for determining the purity of water. This method uses a collection of water quality variables, including pH, turbidity, conductivity, and temperature, to train a model using a classification algorithm. The model's accuracy in determining the purity of the water is next assessed using a new set of data. The outcomes show how effective the suggested method is at assessing the purity of the water. Consequently, this method is a reliable and efficient tool for determining the purity of water, which can significantly lower the danger of water pollution. Monitoring water quality to regulate it, safeguard public health, and save the environment is crucial. With the support of various machine learning classifiers and related stacking ensemble models, the WQI was primarily used to categorize the WQ data. Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), Decision Tree (DT), and Naive Bayes are a few of the classifiers that were used. In total, 3996 samples collected during nine years were included in the dataset used for the study, along with supporting data.