{"id":26246012,"url":"https://github.com/blei-lab/deconfounder_public","last_synced_at":"2025-10-20T08:09:42.412Z","repository":{"id":85552959,"uuid":"276715090","full_name":"blei-lab/deconfounder_public","owner":"blei-lab","description":null,"archived":false,"fork":false,"pushed_at":"2022-12-19T15:53:30.000Z","size":36628,"stargazers_count":21,"open_issues_count":2,"forks_count":6,"subscribers_count":27,"default_branch":"master","last_synced_at":"2025-06-23T20:58:27.350Z","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":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/blei-lab.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE.md","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":"2020-07-02T18:07:56.000Z","updated_at":"2025-04-28T06:27:50.000Z","dependencies_parsed_at":"2023-03-03T15:15:28.079Z","dependency_job_id":null,"html_url":"https://github.com/blei-lab/deconfounder_public","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/blei-lab/deconfounder_public","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/blei-lab%2Fdeconfounder_public","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/blei-lab%2Fdeconfounder_public/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/blei-lab%2Fdeconfounder_public/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/blei-lab%2Fdeconfounder_public/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/blei-lab","download_url":"https://codeload.github.com/blei-lab/deconfounder_public/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/blei-lab%2Fdeconfounder_public/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":262406937,"owners_count":23306282,"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":"2025-03-13T13:17:08.815Z","updated_at":"2025-10-20T08:09:42.272Z","avatar_url":"https://github.com/blei-lab.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Reference implementation of the deconfounder\n\nThis folder contains the code for:\n\n+ the empirical study about smoking (Section 6.1)\n\n+ the empirical study about genome-wide association studies (GWAS)\n  (Section 6.2)\n\n+ the empirical study about movies (Section 6.3)\n\n# Acknowledgments\n\nWe thank Justin Grimmer, Dean Knox, and Brandon Stewart for pointing\nout concerns with a previous version of the code.\n\n+ Smoking study:\n\n\t+ There was a bug with how the code handled heldout data in the\n\t  Stan file for fitting factor models. This bug has been fixed.\n\n\t+ Regarding the definition of bias, variance, and MSE metrics: We\n\t  now have added an explanation of these metrics in\n\t  `utils.R`. These metrics are defined as average per-simulation\n\t  posterior bias/variance/MSE, following Korattikara et\n\t  al. (2014); Chen et al. (2015); Gustafson (2015).\n\n\t+ Regarding the number of posterior samples drawn to compute the\n\t  Monte Carlo estimate of posterior bias/variance/MSE: We have\n\t  increased the default number of posterior samples to 10. (This\n\t  number can be set as a hyperparameter and further increased.)\n\n+ Movie study:\n\n\t+ [Grimmer et al.  (2020)](https://www.dropbox.com/s/71m4ncw6s9nkek7/gks.pdf)\n\t  points out that conditioning on observed covariates can be\n\t  important for causal estimation. Using the movie data, they\n\t  illustrate how the deconfounder without conditioning on observed\n\t  covariates can produce unreasonable causal estimates.\n\n      The exploratory results on actors in Wang and Blei (2019) do not\n\t  condition on covariates.  The \"intervened test set\" analyses in\n\t  the supplement of Wang and Blei (2019) includes such\n\t  conditioning.\n\n      For users that are interested in exploring different analyses,\n\t  the causal estimation section of\n\t  `movie_py/movie-actor-py2-causalest.ipynb` includes a block\n\t  of code that conditions on observed covariates in addition to\n\t  the substitute confounder.\n\n# References\n\nChen, C., Ding, N., \u0026 Carin, L. (2015). On the convergence of\nstochastic gradient MCMC algorithms with high-order integrators. In\nAdvances in Neural Information Processing Systems (pp. 2278-2286).\n\n\nGrimmer, J., Knox, D., \u0026 Stewart, B. (2020). Naive regression requires\nweaker assumptions than factor models to adjust for multiple cause\nconfounding\n[[link](https://www.dropbox.com/s/71m4ncw6s9nkek7/gks.pdf)]\n\n\nGustafson, P. (2015). Bayesian inference for partially identified\nmodels: Exploring the limits of limited data (Vol. 140). CRC Press.\n\n\nKorattikara, A., Chen, Y., \u0026 Welling, M. (2014). Austerity in MCMC\nland: Cutting the Metropolis-Hastings budget. In International\nConference on Machine Learning (pp. 181-189).\n\n\nWang, Y. and Blei, D.M. (2019) The Blessings of Multiple Causes.\n_Journal of American Statistical Association_, 114:528, 1574-1596.\n[[link](https://amstat.tandfonline.com/doi/full/10.1080/01621459.2019.1686987?af=R)]\n\n\n\n\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fblei-lab%2Fdeconfounder_public","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fblei-lab%2Fdeconfounder_public","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fblei-lab%2Fdeconfounder_public/lists"}