AutoDCWorkflow: LLM-based Data Cleaning Workflow Auto-Generation and Benchmark

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Li, Lan, Fang, Liri, Ludäscher, Bertram, Torvik, Vetle I.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911117942456320
author Li, Lan
Fang, Liri
Ludäscher, Bertram
Torvik, Vetle I.
author_facet Li, Lan
Fang, Liri
Ludäscher, Bertram
Torvik, Vetle I.
contents Data cleaning is a time-consuming and error-prone manual process, even with modern workflow tools such as OpenRefine. We present AutoDCWorkflow, an LLM-based pipeline for automatically generating data-cleaning workflows. The pipeline takes a raw table and a data analysis purpose, and generates a sequence of OpenRefine operations designed to produce a minimal, clean table sufficient to address the purpose. Six operations correspond to common data quality issues, including format inconsistencies, type errors, and duplicates. To evaluate AutoDCWorkflow, we create a benchmark with metrics assessing answers, data, and workflow quality for 142 purposes using 96 tables across six topics. The evaluation covers three key dimensions: (1) Purpose Answer: can the cleaned table produce a correct answer? (2) Column (Value): how closely does it match the ground truth table? (3) Workflow (Operations): to what extent does the generated workflow resemble the human-curated ground truth? Experiments show that Llama 3.1, Mistral, and Gemma 2 significantly enhance data quality, outperforming the baseline across all metrics. Gemma 2-27B consistently generates high-quality tables and answers, while Gemma 2-9B excels in producing workflows that closely resemble human-annotated versions.
format Preprint
id arxiv_https___arxiv_org_abs_2412_06724
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AutoDCWorkflow: LLM-based Data Cleaning Workflow Auto-Generation and Benchmark
Li, Lan
Fang, Liri
Ludäscher, Bertram
Torvik, Vetle I.
Databases
Computation and Language
Data cleaning is a time-consuming and error-prone manual process, even with modern workflow tools such as OpenRefine. We present AutoDCWorkflow, an LLM-based pipeline for automatically generating data-cleaning workflows. The pipeline takes a raw table and a data analysis purpose, and generates a sequence of OpenRefine operations designed to produce a minimal, clean table sufficient to address the purpose. Six operations correspond to common data quality issues, including format inconsistencies, type errors, and duplicates. To evaluate AutoDCWorkflow, we create a benchmark with metrics assessing answers, data, and workflow quality for 142 purposes using 96 tables across six topics. The evaluation covers three key dimensions: (1) Purpose Answer: can the cleaned table produce a correct answer? (2) Column (Value): how closely does it match the ground truth table? (3) Workflow (Operations): to what extent does the generated workflow resemble the human-curated ground truth? Experiments show that Llama 3.1, Mistral, and Gemma 2 significantly enhance data quality, outperforming the baseline across all metrics. Gemma 2-27B consistently generates high-quality tables and answers, while Gemma 2-9B excels in producing workflows that closely resemble human-annotated versions.
title AutoDCWorkflow: LLM-based Data Cleaning Workflow Auto-Generation and Benchmark
topic Databases
Computation and Language
url https://arxiv.org/abs/2412.06724