Imputing missing ground water level data with deep learning
We used a hierarchical hybrid model to impute missing ground water level time series data.
Project Duration: 2023-Present
Highlights
- We proposed a hybrid model, comprising 1D convolutions and stacked LSTM layers, to impute missing ground water levels.
- Our model has shown promising results so far, achieving an MAE of 0.20 for 25% missing data.