{"id":13571412,"url":"https://github.com/jsyoon0823/TimeGAN","last_synced_at":"2025-04-04T08:31:10.742Z","repository":{"id":41472864,"uuid":"258629115","full_name":"jsyoon0823/TimeGAN","owner":"jsyoon0823","description":"Codebase for Time-series Generative Adversarial Networks (TimeGAN) - NeurIPS 2019","archived":false,"fork":false,"pushed_at":"2024-07-30T21:00:20.000Z","size":1312,"stargazers_count":855,"open_issues_count":1,"forks_count":270,"subscribers_count":10,"default_branch":"master","last_synced_at":"2024-11-05T04:34:09.230Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jsyoon0823.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"license","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2020-04-24T21:32:32.000Z","updated_at":"2024-11-04T00:22:24.000Z","dependencies_parsed_at":"2023-02-12T12:45:53.053Z","dependency_job_id":"0b15d2eb-8809-454a-8f73-283724e1987f","html_url":"https://github.com/jsyoon0823/TimeGAN","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jsyoon0823%2FTimeGAN","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jsyoon0823%2FTimeGAN/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jsyoon0823%2FTimeGAN/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jsyoon0823%2FTimeGAN/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jsyoon0823","download_url":"https://codeload.github.com/jsyoon0823/TimeGAN/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247147095,"owners_count":20891623,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-08-01T14:01:01.794Z","updated_at":"2025-04-04T08:31:08.976Z","avatar_url":"https://github.com/jsyoon0823.png","language":"Jupyter Notebook","funding_links":[],"categories":["Libraries","📝 Papers with code","时间序列"],"sub_categories":["Managed database services","网络服务_其他"],"readme":"# Codebase for \"Time-series Generative Adversarial Networks (TimeGAN)\"\n\nAuthors: Jinsung Yoon, Daniel Jarrett, Mihaela van der Schaar\n\nReference: Jinsung Yoon, Daniel Jarrett, Mihaela van der Schaar, \n\"Time-series Generative Adversarial Networks,\" \nNeural Information Processing Systems (NeurIPS), 2019.\n \nPaper Link: https://papers.nips.cc/paper/8789-time-series-generative-adversarial-networks\n\nContact: jsyoon0823@gmail.com\n\nThis directory contains implementations of TimeGAN framework for synthetic time-series data generation\nusing one synthetic dataset and two real-world datasets.\n\n-   Sine data: Synthetic\n-   Stock data: https://finance.yahoo.com/quote/GOOG/history?p=GOOG\n-   Energy data: http://archive.ics.uci.edu/ml/datasets/Appliances+energy+prediction\n\nTo run the pipeline for training and evaluation on TimeGAN framwork, simply run \npython3 -m main_timegan.py or see jupyter-notebook tutorial of TimeGAN in tutorial_timegan.ipynb.\n\nNote that any model architecture can be used as the generator and \ndiscriminator model such as RNNs or Transformers. \n\n### Code explanation\n\n(1) data_loading.py\n- Transform raw time-series data to preprocessed time-series data (Googld data)\n- Generate Sine data\n\n(2) Metrics directory\n  (a) visualization_metrics.py\n  - PCA and t-SNE analysis between Original data and Synthetic data\n  (b) discriminative_metrics.py\n  - Use Post-hoc RNN to classify Original data and Synthetic data\n  (c) predictive_metrics.py\n  - Use Post-hoc RNN to predict one-step ahead (last feature)\n\n(3) timegan.py\n- Use original time-series data as training set to generater synthetic time-series data\n\n(4) main_timegan.py\n- Report discriminative and predictive scores for the dataset and t-SNE and PCA analysis\n\n(5) utils.py\n- Some utility functions for metrics and timeGAN.\n\n### Command inputs:\n\n-   data_name: sine, stock, or energy\n-   seq_len: sequence length\n-   module: gru, lstm, or lstmLN\n-   hidden_dim: hidden dimensions\n-   num_layers: number of layers\n-   iterations: number of training iterations\n-   batch_size: the number of samples in each batch\n-   metric_iterations: number of iterations for metric computation\n\nNote that network parameters should be optimized for different datasets.\n\n### Example command\n\n```shell\n$ python3 main_timegan.py --data_name stock --seq_len 24 --module gru\n--hidden_dim 24 --num_layer 3 --iteration 50000 --batch_size 128 \n--metric_iteration 10\n```\n\n### Outputs\n\n-   ori_data: original data\n-   generated_data: generated synthetic data\n-   metric_results: discriminative and predictive scores\n-   visualization: PCA and tSNE analysis","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjsyoon0823%2FTimeGAN","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjsyoon0823%2FTimeGAN","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjsyoon0823%2FTimeGAN/lists"}