Scholarlink

<p>Water is essential to human health, economic development, and productivity. Many contaminants have significantly affected the quality of water over time; thus, predicting water quality using modern tools is vital for mitigating water pollution. This research focused on an intensive examination of two distinct lakes. 80 surface water samples were taken randomly from each lake for 8&#xa0;months continuously and tested in laboratory for physicochemical characteristics to determine the drinkability of the two lakes. However, 8 Spin Karez samples Lake and 18 Hanna Lake samples were deemed unsafe to drink.&#xa0;The samples were classified as drinkable or non-drinkable based on their drinkability values. The first six months of drinkability data were used to train the algorithms, and then the remaining two months of drinkability data were forecasted. Confusion matrix was used to examine the prediction performance of seven distinct supervised machine learning (ML) algorithms – Linear Regression (LR), Decision Tree Classifier (DTC), Random Forest (RF), XGBoosting (XGB), K-Nearest Neighbor (KNN), Support Vector Machines (SVM), and Adaptive Boosting (Adaboost). The models accurately predicted the water quality with DTC outperforming for both lakes. This study contributes to the evolution of prediction tools by integrating ML with environmental monitoring for sustainable water resource management.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predictive Machine Learning for Seasonal Water Quality Insights

  • 期刊论文

  • 发表日期：

    2026年1月21日

查看全文
  • Zouhaib Ali,
  • Muhammad Bilal,
  • Sallahuddin Panhwar,
  • Muhammad Junaid,
  • Hareef Ahmed Keerio,
  • Israr Hussain

摘要

Water is essential to human health, economic development, and productivity. Many contaminants have significantly affected the quality of water over time; thus, predicting water quality using modern tools is vital for mitigating water pollution. This research focused on an intensive examination of two distinct lakes. 80 surface water samples were taken randomly from each lake for 8 months continuously and tested in laboratory for physicochemical characteristics to determine the drinkability of the two lakes. However, 8 Spin Karez samples Lake and 18 Hanna Lake samples were deemed unsafe to drink. The samples were classified as drinkable or non-drinkable based on their drinkability values. The first six months of drinkability data were used to train the algorithms, and then the remaining two months of drinkability data were forecasted. Confusion matrix was used to examine the prediction performance of seven distinct supervised machine learning (ML) algorithms – Linear Regression (LR), Decision Tree Classifier (DTC), Random Forest (RF), XGBoosting (XGB), K-Nearest Neighbor (KNN), Support Vector Machines (SVM), and Adaptive Boosting (Adaboost). The models accurately predicted the water quality with DTC outperforming for both lakes. This study contributes to the evolution of prediction tools by integrating ML with environmental monitoring for sustainable water resource management.

Graphical Abstract