Pump It Up: Predict Water Pump Status using Attentive Tabular Learning
This work addresses timely maintenance of water pumps in drought-hit communities, but it is incremental as it applies an existing method (TabNet) to a specific dataset with a minor modification (focal loss).
The paper tackles predicting water pump repair status in Tanzania using TabNet, a sequential attentive deep neural architecture, and shows performance improvements by using focal loss on imbalanced data, achieving competitive results compared to gradient tree-boosting algorithms like XGBoost, LightGBM, and CatBoost.
Water crisis is a crucial concern around the globe. Appropriate and timely maintenance of water pumps in drought-hit countries is vital for communities relying on the well. In this paper, we analyze and apply a sequential attentive deep neural architecture, TabNet, for predicting water pump repair status in Tanzania. The model combines the valuable benefits of tree-based algorithms and neural networks, enabling end-to-end training, model interpretability, sparse feature selection, and efficient learning on tabular data. Finally, we compare the performance of TabNet with popular gradient tree-boosting algorithms like XGBoost, LightGBM,CatBoost, and demonstrate how we can further uplift the performance by choosing focal loss as the objective function while training on imbalanced data.