{"id":28731667,"url":"https://github.com/andrewdarnall/liquidsky","last_synced_at":"2025-06-15T18:40:36.171Z","repository":{"id":298763152,"uuid":"870645714","full_name":"AndrewDarnall/LiquidSky","owner":"AndrewDarnall","description":"A benchmarking of univariate time series forecasting with statisticsal models and deep learning models","archived":false,"fork":false,"pushed_at":"2025-06-12T19:35:51.000Z","size":7902,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-06-12T20:53:28.981Z","etag":null,"topics":["arima","data-analytics","liquid-network","lstm","time-series-forecasting","transformer"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/AndrewDarnall.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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,"zenodo":null}},"created_at":"2024-10-10T12:17:32.000Z","updated_at":"2025-06-12T19:36:15.000Z","dependencies_parsed_at":"2025-06-12T21:06:15.526Z","dependency_job_id":null,"html_url":"https://github.com/AndrewDarnall/LiquidSky","commit_stats":null,"previous_names":["andrewdarnall/tsf-benchmark"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/AndrewDarnall/LiquidSky","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AndrewDarnall%2FLiquidSky","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AndrewDarnall%2FLiquidSky/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AndrewDarnall%2FLiquidSky/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AndrewDarnall%2FLiquidSky/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/AndrewDarnall","download_url":"https://codeload.github.com/AndrewDarnall/LiquidSky/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AndrewDarnall%2FLiquidSky/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":260029959,"owners_count":22948253,"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":["arima","data-analytics","liquid-network","lstm","time-series-forecasting","transformer"],"created_at":"2025-06-15T18:40:31.166Z","updated_at":"2025-06-15T18:40:36.158Z","avatar_url":"https://github.com/AndrewDarnall.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Liquid Sky - A Time Series Forecasting Benchmark\n\n![TSF-Benchmark Cover](./assets/imgs/TSF-Benchmark-Cover.png)\n\n---\n\nThis project explores the benchmarking of **univariate time series forecasting**, focusing on comparing the performances of various models. The models analyzed include traditional \n**statistical models** such as **ARIMA** and **SARIMA**, as well as modern **deep learning models** like **LSTM** and **Transformer** networks. \nAdditionally, the study delves into a new class of emerging technologies, **continuous deep learning models**, \nwith a particular emphasis on **Liquid Time Constant Networks** and their **Closed-form Approximation**.\n\nThe benchmark is conducted using a simple univariate time series dataset, serving as a reference for all models. \nBeyond evaluating the **inference performance** of each model, the analysis also considers the **advantages and trade-offs** \nof using one class of models over others for similar forecasting tasks. \nThis comprehensive approach provides insights into the conditions under which certain models may outperform others, \nfacilitating more informed decision-making in model selection for time series forecasting.\n\n\n------\n\n# Requirements\n\n| Component | Version |\n|-----------|---------|\n| Python    | `3.12.5`|\n| pip       | `22.0.4`|\n\n\n# Set Up the Python Environment\n\n```bash\npython -m venv data-analytics\n```\n\n# Download the Dependencies\n\n```bash\npython -m pip install -r requirements.txt\n```\n------\n\n# Download the Datasets and Run the notebook\n\nClone the GitHub repo and change into the repo's directory:\n\n```bash\ngit clone https://github.com/AndrewDarnall/TSF-Benchmark.git\ncd TSF-Benchmark\n```\n\nCreate the datasets directory:\n\n```bash\nmkdir datasets\ncd datasets\n```\n\nThe following notebook uses the [AirPassengers](https://www.kaggle.com/datasets/rakannimer/air-passengers) dataset, which once downloaded, extract the .zip file into the `datasets/` directory\n\n```bash\nmv ~/Downloads/archive.zip .\nunzip archive.zip\nrm archive.zip\ncd ..\n```\n\nFrom the commandline terminal (in Linux and macOS) run the following command:\n\n```bash\njupyter-notebook ./TSF-Benchmark.ipynb --port=8086\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fandrewdarnall%2Fliquidsky","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fandrewdarnall%2Fliquidsky","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fandrewdarnall%2Fliquidsky/lists"}