Scholarlink

<p>This study proposes and evaluates curve-fitting models—specifically polynomial and Gaussian techniques—to accurately compute the pH sub-index as part of a Water Quality Index (WQI). This modeling is essential for enhancing the reliability of pH representation in real-time Internet of Things (IoT)-based water quality monitoring systems. By improving the pH sub-index formulation, the research aims to address one of the core components in WQI computations under dynamic environmental conditions. This work represents a critical step in the second phase of developing the IoT-based Water Quality Index (IoTWQI) framework. The resulting models are validated and compared for their suitability in sensor-based environments. The findings highlight the Gaussian Model 1 as the most accurate choice for normally distributed data (RMSE = 2.15), whereas the 6th-degree Polynomial Model remains effective in capturing complex nonlinear variations (RMSE = 4.26). These insights contribute directly to improving real-time WQI calculation performance in IoT ecosystems.</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

Evaluating polynomial and Gaussian fitting techniques for accurate sub-index pH modeling in water quality index development

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月9日

查看全文
  • Wibowo Harry Sugiharto,
  • Agung Budi Prasetijo,
  • Heru Susanto

摘要

This study proposes and evaluates curve-fitting models—specifically polynomial and Gaussian techniques—to accurately compute the pH sub-index as part of a Water Quality Index (WQI). This modeling is essential for enhancing the reliability of pH representation in real-time Internet of Things (IoT)-based water quality monitoring systems. By improving the pH sub-index formulation, the research aims to address one of the core components in WQI computations under dynamic environmental conditions. This work represents a critical step in the second phase of developing the IoT-based Water Quality Index (IoTWQI) framework. The resulting models are validated and compared for their suitability in sensor-based environments. The findings highlight the Gaussian Model 1 as the most accurate choice for normally distributed data (RMSE = 2.15), whereas the 6th-degree Polynomial Model remains effective in capturing complex nonlinear variations (RMSE = 4.26). These insights contribute directly to improving real-time WQI calculation performance in IoT ecosystems.