International Research Journal of Education and Technology

Accurate prediction of surface water quality plays a vital role in environmental monitoring and ensuring public health. Traditionally, manual sampling and statistical approaches have been employed for this purpose; however, these methods are time-intensive and often fail to capture intricate patterns present in water quality data. This study leverages an LSTM-based deep learning model to enhance the precision of the Water Quality Index (WQI) prediction. Surface water quality is affected by factors such as seasonal fluctuations, pollution incidents, and climatic variations. Unlike conventional models that primarily account for short-term relationships, LSTM effectively captures long-term dependencies through its gating mechanisms (forget, input, and output), enabling it to emphasize significant trends while minimizing noise. Missing values in the dataset are addressed using mean imputation, and data preprocessing is carried out with MinMaxScaler for feature normalization. The model’s performance is evaluated using the R² score, achieving a high accuracy of 99.9%, demonstrating the effectiveness of this approach.

Keywords:LSTM; MinMax Scaler; R² Score;Surface Water; WQI.