Enhanced Water Quality Index Prediction through Machine Learning and IoT with Hybrid Voting Methods

Le résumé fourni par la source

Abstract Water is a vital resource for our daily existence and plays a crucial role. Therefore, utilizing artificial intelligence and Internet of Things (IoT) devices can be imperative to monitor water quality efficiently. To address water quality, this study presents the development of a hybrid voting method for detecting water quality. The objectives of this study: 1) develop a hybrid voting model; 2) Analyze the ML model’s memory and time complexity; 3) Integrate the best model with an IoT to monitor water quality. However, the voting method, consisting of LR, SVM, and XGBoost models, achieved an accuracy of 96%. Furthermore, we extensively examined diverse machine learning models, encompassing both instances with and without feature reduction. The analysis reveals that the models, namely XGBoost, SVM, LR, KNN, RF, and DT, exhibited accuracy rates of 94, 95, 85, 92, and 90%, respectively, without employing any feature reduction techniques. In contrast, when utilizing the PCA feature reduction technique, the accuracy percentages for the same models were 89, 92, 91, 84, 86 and 78%, respectively. Ultimately, we thoroughly examined our suggested model, considering variables such as memory complexity, time complexity, statistical error analysis, measurements based on the Internet of Things (IoT), and a comparative investigation with other research discoveries.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.