Datasets/Tg/readme.md

Data Sources:

• data/train.csv: https://www.kaggle.com/competitions/neurips-open-polymer-prediction-2025/data?select=train.csv

• data/train_supplement/dataset3.csv: https://www.kaggle.com/competitions/neurips-open-polymer-prediction-2025/data

• Tg_SMILES_class_pid_polyinfo_median.csv: https://github.com/Yeaaahhhhh/polyerdatasets/blob/main/Tg_SMILES_class_pid_polyinfo_median.csv

• tg_data_2: https://springernature.figshare.com/articles/dataset/dataset_with_glass_transition_temperature/24219958?file=42507037

• tg_data_1: https://www.sciencedirect.com/science/article/pii/S2590159123000377#ec0005

• pi.csv: https://github.com/hkqiu/DataAugmentation4SmallData/blob/main/model/training%20data/pi.csv

• polymer.csv: https://github.com/hkqiu/DataAugmentation4SmallData/blob/main/model/training%20data/polymer.csv

• tg_density.csv: https://www.kaggle.com/datasets/oleggromov/polymer-tg-density-excerpt || https://github.com/Shorku/rhnet/tree/main/data

Steps:

• Download the above datasets.

• Run the Datasets/Tg/data_augmentation.ipynb file to get the Tg.csv augmented data.

• Or Use the already available Tg.csv.

• Use the NIPS GNN for Training & Inference on Tg Property.