Scholarlink

<p>Missing values in tabular data lakes can severely impact data analysis and diminish the performance in downstream applications. We highlight that a robust imputation strategy should properly take three aspects of variety into consideration: source of imputed value, the types of tables involved, and the data types of the missing value. Existing imputation methods rely on estimation-based approaches (using a model trained on data from the same table to estimate missing values) or search-based approaches (retrieving values from other tables). Unfortunately, none of these approaches effectively incorporate all three aspects of variety. To address this gap, we propose <Emphasis FontCategory="SansSerif">CESID</Emphasis>, a novel framework that uses a <Emphasis Type="Underline">C</Emphasis>ombination of <Emphasis Type="Underline">E</Emphasis>stimation-based and <Emphasis Type="Underline">S</Emphasis>earch-based methods for missing value <Emphasis Type="Underline">I</Emphasis>mputation in <Emphasis Type="Underline">D</Emphasis>ata lakes. <Emphasis FontCategory="SansSerif">CESID</Emphasis> contains three core modules: (1) the <Emphasis FontCategory="SansSerif">Contextual Search Module</Emphasis>, which efficiently discovers candidate values from tables by exploiting the contextual information; (2) the <Emphasis FontCategory="SansSerif">Acquisition-guided Estimation Module</Emphasis>, which introduces an influence function and a sampling-based exploration strategy to yield accurate estimated values; (3) the <Emphasis FontCategory="SansSerif">Classifier Module</Emphasis>, which determines the most suitable method based on table-level and column-level statistics. Extensive experiments conducted on three data lakes demonstrate that <Emphasis FontCategory="SansSerif">CESID</Emphasis> effectively and efficiently addresses the missing value problem.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Missing Value Imputation in Tabular Data Lakes Unleashed: A Hybrid Approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月19日

查看全文
  • Feng Luo,
  • Hai Lan,
  • Hui Luo,
  • Zhifeng Bao,
  • J. Shane Culpepper,
  • Shazia Sadiq,
  • Xiaoli Wang

摘要

Missing values in tabular data lakes can severely impact data analysis and diminish the performance in downstream applications. We highlight that a robust imputation strategy should properly take three aspects of variety into consideration: source of imputed value, the types of tables involved, and the data types of the missing value. Existing imputation methods rely on estimation-based approaches (using a model trained on data from the same table to estimate missing values) or search-based approaches (retrieving values from other tables). Unfortunately, none of these approaches effectively incorporate all three aspects of variety. To address this gap, we propose CESID, a novel framework that uses a Combination of Estimation-based and Search-based methods for missing value Imputation in Data lakes. CESID contains three core modules: (1) the Contextual Search Module, which efficiently discovers candidate values from tables by exploiting the contextual information; (2) the Acquisition-guided Estimation Module, which introduces an influence function and a sampling-based exploration strategy to yield accurate estimated values; (3) the Classifier Module, which determines the most suitable method based on table-level and column-level statistics. Extensive experiments conducted on three data lakes demonstrate that CESID effectively and efficiently addresses the missing value problem.