首页|Studies from Durban University of Technology Yield New Data on Machine Learning (Prediction of Wastewater Quality Parameters Using Adaptive and Machine Learning Models: a South African Case Study)

Studies from Durban University of Technology Yield New Data on Machine Learning (Prediction of Wastewater Quality Parameters Using Adaptive and Machine Learning Models: a South African Case Study)

扫码查看
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News-Investigators discuss new findings in Machine Learning. According to news reportingoriginating from Durban, South Afr ica, by NewsRx correspondents, research stated, "The wastewatertreatment proces s often faces challenges in monitoring water quality parameters (WQ), to overcom e thisthere is a need for developing innovative modeling approaches. Hence, the present study is motivatedby the potential application of adaptive and machine learning (ML) models as soft sensors to predict theWQ in one of the largest Mu nicipal Wastewater Treatment Plants (MWWTP) in KwaZulu-Natal, SouthAfrica."Financial support for this research came from National Research Foundation - Sou th Africa.Our news editors obtained a quote from the research from the Durban University o f Technology, "Sevendifferent adaptive and ML algorithms were examined and comp ared, varying from adaptive strategies to MLarchitectures such as Long Short-Te rm Memory (LSTM), Bidirectional LSTM (BiLSTM), Time Difference(TD), Just in Tim e Learning (JIT), Moving Window (MW), and fusion of adaptive strategies (JITTD,and JITTDMW), Support Vector Regression (SVR), and Artificial Neural Network (AN N). Based on theresults, BiLSTM consistently provided the most accurate estimat ion of effluent parameters, with an errorrate ranging from 3.12 to 9.75 % for all variables. For Chemical Oxygen Demand (COD), ammonia, pH,and Total Susp ended Solids (TSS), BiLSTM model yielded low errors (Mean Absolute Error (MAE) values of 1.54, 0.1, 0.22, and 1.14) with lower correlation coefficient values (<0.7) compared to the sixother models proposed. As for conductivity, COD, TSS, p H, ammonia, LSTM, and JITTDMW, JITTDperformed well with MAE values between 1 an d 8 but had difficulty estimating soluble reactive phosphate(SRP). From a futur e perspective, these models could be applied to other MWWTPs facing similar challenges, potentially helping to improve their performance and effectiveness."

DurbanSouth AfricaAfricaCyborgsE merging TechnologiesMachine LearningDurban University of Technology

2024

Robotics & Machine Learning Daily News

Robotics & Machine Learning Daily News

ISSN:
年,卷(期):2024.(Oct.31)
  • 1