Home

Awesome

TSEst

TSEst is a multi-modal time series imputation model that can use an additional modality (another cross-sectional or time-series data) to impute missing values in a time series data. This multi-modal approach shows improved performance over uni-modal imputation models.

Framework

<img src="https://github.com/compbiolabucf/TSEst/blob/main/Fig-1.png" width="450" height="450">

Environment

Environment can be created using the command conda env create -f my_conda_env.yml. my_conda_env.yml is provided in the repository.

Quick start guide

Download a sample data (Daymet) from this link into the parent directory. Run python3 run_models.py --config_path configs/Camel_Transformer_best_rnd.ini to train the model. Modify the values of <model_saving_dir> and