Scholarlink

For accurate pH measurement and voltage prediction, solution properties need to be converted into measurable electrical signals. Reliable and precise data from pH sensors are achieved by calibrating the sensors with known solutions, supporting effective monitoring and maintenance of water quality to safeguard organism health in aquaculture applications. In this study, fabricated pH strips are used for data measurement, and a dataset spanning various pH levels has been developed. Using machine learning algorithms, voltage is predicted across these pH values, with random forest regression yielding the best prediction accuracy for different pH ranges. This approach enhances sensor precision and reliability, minimizing the need for manual calibration. Additionally, it enables real-time pH monitoring and effectively captures complex parameter relationships, offering a more efficient and sustainable solution for optimal water quality management in aquaculture. The random forest model worked remarkably well, with a low RMSE of 0.01V and a R2 value of 0.99, indicating great accuracy and minimum error in voltage predictions. The high R2 score indicates that the model explains nearly all of the variance in the data.

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

Machine Learning Based Voltage Prediction with Various Levels of pH in IoT Environment

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • G. R. Kanagachidambaresan,
  • P. Rohini,
  • V. Jayasudha,
  • S. P. Thipperudraswamy,
  • Pandiaraj Manickam,
  • J. Mathiyarasu,
  • P. Kumararaja,
  • M. Muralidhar

摘要

For accurate pH measurement and voltage prediction, solution properties need to be converted into measurable electrical signals. Reliable and precise data from pH sensors are achieved by calibrating the sensors with known solutions, supporting effective monitoring and maintenance of water quality to safeguard organism health in aquaculture applications. In this study, fabricated pH strips are used for data measurement, and a dataset spanning various pH levels has been developed. Using machine learning algorithms, voltage is predicted across these pH values, with random forest regression yielding the best prediction accuracy for different pH ranges. This approach enhances sensor precision and reliability, minimizing the need for manual calibration. Additionally, it enables real-time pH monitoring and effectively captures complex parameter relationships, offering a more efficient and sustainable solution for optimal water quality management in aquaculture. The random forest model worked remarkably well, with a low RMSE of 0.01V and a R2 value of 0.99, indicating great accuracy and minimum error in voltage predictions. The high R2 score indicates that the model explains nearly all of the variance in the data.