<![CDATA[Efficient water quality prediction is crucial amid the threat of increasing pollution. Conventional methods have limitations in terms of cost, time, and coverage, requiring an innovative approach based on artificial intelligence. This study aims to classify drinking water suitability using Long Short-Term Memory (LSTM) architecture, which is known to be effective for sequential data. The research method includes data collection from Kaggle, data pre-processing such as normalization and missing value handling, and data division into 80% for training and 20% for testing. The proposed LSTM model was evaluated using accuracy, precision, recall, and F1-Score metrics. The results show that the LSTM model is capable of achieving an overall accuracy of 95.47%, with a precision of 0.9511 and a recall of 0.9547. Although there are some False Positive classification errors (predicting water as unfit when it is actually fit), the overall performance of the model is excellent and reliable for this classification task. The conclusion of this study is that the LSTM model can be an effective and accurate solution for predicting water quality, supporting early detection and real-time pollution control efforts.]]>
Copyrights © 2026
