HoloClean : Holistic Data Repairs with Probabilistic Inference. Rekatsinas, T., Chu, X., Ilyas, I., F., & Ré, C. 10(i):1-13, 2017. Paper abstract bibtex We introduce HoloClean, a framework for holistic data repairing driven by probabilistic inference. HoloClean unifies existing quali-tative data repairing approaches, which rely on integrity constraints or external data sources, with quantitative data repairing methods, which leverage statistical properties of the input data. Given an inconsistent dataset as input, HoloClean automatically generates a probabilistic program that performs data repairing. Inspired by re-cent theoretical advances in probabilistic inference, we introduce a series of optimizations which ensure that inference over Holo-Clean's probabilistic model scales to instances with millions of tu-ples. We show that HoloClean scales to instances with millions of tuples and find data repairs with an average precision of ∼ 90% and an average recall of above ∼ 76% across a diverse array of datasets exhibiting different types of errors. This yields an average F1 improvement of more than 2× against state-of-the-art methods.
@article{
title = {HoloClean : Holistic Data Repairs with Probabilistic Inference},
type = {article},
year = {2017},
identifiers = {[object Object]},
pages = {1-13},
volume = {10},
id = {c10f130d-b83d-3dba-9c24-e3eaecade36d},
created = {2017-08-29T16:55:16.589Z},
file_attached = {true},
profile_id = {9d42c173-db8a-3566-8a55-b1a79270b5ed},
group_id = {325c0157-5788-3a71-bd59-340f23437b88},
last_modified = {2018-01-10T17:12:34.919Z},
read = {false},
starred = {false},
authored = {false},
confirmed = {true},
hidden = {false},
citation_key = {Rekatsinas2017},
private_publication = {false},
abstract = {We introduce HoloClean, a framework for holistic data repairing driven by probabilistic inference. HoloClean unifies existing quali-tative data repairing approaches, which rely on integrity constraints or external data sources, with quantitative data repairing methods, which leverage statistical properties of the input data. Given an inconsistent dataset as input, HoloClean automatically generates a probabilistic program that performs data repairing. Inspired by re-cent theoretical advances in probabilistic inference, we introduce a series of optimizations which ensure that inference over Holo-Clean's probabilistic model scales to instances with millions of tu-ples. We show that HoloClean scales to instances with millions of tuples and find data repairs with an average precision of ∼ 90% and an average recall of above ∼ 76% across a diverse array of datasets exhibiting different types of errors. This yields an average F1 improvement of more than 2× against state-of-the-art methods.},
bibtype = {article},
author = {Rekatsinas, Theodoros and Chu, Xu and Ilyas, Ihab F and Ré, Christopher},
number = {i}
}
Downloads: 0
{"_id":"vyX3QfdSpHpFKuMGh","bibbaseid":"rekatsinas-chu-ilyas-r-holocleanholisticdatarepairswithprobabilisticinference-2017","authorIDs":[],"author_short":["Rekatsinas, T.","Chu, X.","Ilyas, I., F.","Ré, C."],"bibdata":{"title":"HoloClean : Holistic Data Repairs with Probabilistic Inference","type":"article","year":"2017","identifiers":"[object Object]","pages":"1-13","volume":"10","id":"c10f130d-b83d-3dba-9c24-e3eaecade36d","created":"2017-08-29T16:55:16.589Z","file_attached":"true","profile_id":"9d42c173-db8a-3566-8a55-b1a79270b5ed","group_id":"325c0157-5788-3a71-bd59-340f23437b88","last_modified":"2018-01-10T17:12:34.919Z","read":false,"starred":false,"authored":false,"confirmed":"true","hidden":false,"citation_key":"Rekatsinas2017","private_publication":false,"abstract":"We introduce HoloClean, a framework for holistic data repairing driven by probabilistic inference. HoloClean unifies existing quali-tative data repairing approaches, which rely on integrity constraints or external data sources, with quantitative data repairing methods, which leverage statistical properties of the input data. Given an inconsistent dataset as input, HoloClean automatically generates a probabilistic program that performs data repairing. Inspired by re-cent theoretical advances in probabilistic inference, we introduce a series of optimizations which ensure that inference over Holo-Clean's probabilistic model scales to instances with millions of tu-ples. We show that HoloClean scales to instances with millions of tuples and find data repairs with an average precision of ∼ 90% and an average recall of above ∼ 76% across a diverse array of datasets exhibiting different types of errors. This yields an average F1 improvement of more than 2× against state-of-the-art methods.","bibtype":"article","author":"Rekatsinas, Theodoros and Chu, Xu and Ilyas, Ihab F and Ré, Christopher","number":"i","bibtex":"@article{\n title = {HoloClean : Holistic Data Repairs with Probabilistic Inference},\n type = {article},\n year = {2017},\n identifiers = {[object Object]},\n pages = {1-13},\n volume = {10},\n id = {c10f130d-b83d-3dba-9c24-e3eaecade36d},\n created = {2017-08-29T16:55:16.589Z},\n file_attached = {true},\n profile_id = {9d42c173-db8a-3566-8a55-b1a79270b5ed},\n group_id = {325c0157-5788-3a71-bd59-340f23437b88},\n last_modified = {2018-01-10T17:12:34.919Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {true},\n hidden = {false},\n citation_key = {Rekatsinas2017},\n private_publication = {false},\n abstract = {We introduce HoloClean, a framework for holistic data repairing driven by probabilistic inference. HoloClean unifies existing quali-tative data repairing approaches, which rely on integrity constraints or external data sources, with quantitative data repairing methods, which leverage statistical properties of the input data. Given an inconsistent dataset as input, HoloClean automatically generates a probabilistic program that performs data repairing. Inspired by re-cent theoretical advances in probabilistic inference, we introduce a series of optimizations which ensure that inference over Holo-Clean's probabilistic model scales to instances with millions of tu-ples. We show that HoloClean scales to instances with millions of tuples and find data repairs with an average precision of ∼ 90% and an average recall of above ∼ 76% across a diverse array of datasets exhibiting different types of errors. This yields an average F1 improvement of more than 2× against state-of-the-art methods.},\n bibtype = {article},\n author = {Rekatsinas, Theodoros and Chu, Xu and Ilyas, Ihab F and Ré, Christopher},\n number = {i}\n}","author_short":["Rekatsinas, T.","Chu, X.","Ilyas, I., F.","Ré, C."],"urls":{"Paper":"https://bibbase.org/service/mendeley/7127e9bf-bb7e-31a5-a06e-59fccf53ff81/file/52fa12a8-ad40-ab4f-7d9c-47597ec4da81/p1190_rekatsinas.pdf.pdf"},"bibbaseid":"rekatsinas-chu-ilyas-r-holocleanholisticdatarepairswithprobabilisticinference-2017","role":"author","downloads":0},"bibtype":"article","creationDate":"2020-06-19T09:38:54.166Z","downloads":0,"keywords":[],"search_terms":["holoclean","holistic","data","repairs","probabilistic","inference","rekatsinas","chu","ilyas","ré"],"title":"HoloClean : Holistic Data Repairs with Probabilistic Inference","year":2017}