World bank data project instruction

Set up the python env

• conda env create -f environment.yml -- pytorch need to be manually installed with device preference (cuda/cpu)
• conda avtivate wbd

Hyperparameters setting and operations.

1. Task 1: Data Exploration and Missing Value Imputation

   • Hyperparameters: None
   • Run src/task1.py

2. Task 2: Dimensionality Reduction and Clustering using Autoencoder

   • Hyperparameters:
     • seed
     • batch_sizes
     • epochs_list
     • learning_rates
     • dropout_options
     • latent_list
   • Run src/task2.py

3. Task 3: GDP Classification Using MLP

   • Hyperparameters:
     • seed
     • batch_sizes
     • epochs_list
     • learning_rates
     • scoring_metric
     • k_folds
   • Run src/task3.py

4. Task 4: Time-Series GDP Forecasting Using Deep Learning Models

   • Hyperparameters:
     • seed
     • models
     • batch_sizes
     • epochs_list
     • learning_rates
     • dropout_options
     • val_size
     • test_size
   • Run src/task4.py

5. Task 5: Variational Autoencoder for Data Augmentation

   • Hyperparameters:
     • seed
     • batch_size
     • epochs
     • learning_rate
     • hidden_dim
     • latent_dim
     • dataset_timestamp
   • Run src/task5.py

6. Task 6: GDP Forecasting with VAE-Augmented Data

   • Hyperparameters:
     • seed
     • models
     • batch_sizes
     • epochs_list
     • learning_rates
     • dropout_options
     • hidden_dim
     • latent_dim
     • timestamp_task4
     • timestamp_task5
   • Run src/task6.py

Collaborators: Tan Beng Seh and Ng Rou Yan