An AI-based approach to the prediction of water points quality indicators for schistosomiasis prevention

Teegwende Zougmore, Bamba Gueye, Sadouanouan MALO

Keywords: schistosomiasis, water quality prediction, machine learning

Abstract:

We investigate the simultaneous daily forecasting of pH, temperature, dissolved oxygen, and electrical conductivity using AI-based methods. These physicochemical parameters can be retrieved from surface water and favor the reproduction of parasitic worms responsible for Schistosomiasis. Wavelet Artificial Neural Network ($WANN$), Long Short Term Memory ($LSTM$), and Support Vector Regression ($SVR$) are used AI-based methods to build models with fifteen months of collected raw datasets.
They are evaluated through two metrics, such as root-mean-square ($RMSE$) and mean absolute error ($MAE$). The built models take as inputs the physicochemical parameters values observed the last two days and provide as outputs the physicochemical parameters values expected the next day. Overall, the results show that the three methods perform well. The most efficient according to the metrics is the WANN-based model which shows a RMSE of $0.07$, $0.13$, $0.09$, and $9.79$ in forecasting respectively pH, temperature, dissolved oxygen, and electrical conductivity.
and electrical conductivity.

Submission Category: Machine learning algorithms

Submission Number: 68