Table Transformers for Imputing Textual Attributes
This addresses a specific gap in missing data imputation for textual attributes in tabular data, which is incremental as it extends existing methods to handle text.
The paper tackles the problem of missing textual data in tabular datasets by proposing Table Transformers for Imputing Textual Attributes (TTITA), a novel end-to-end transformer-based method that imputes unstructured text columns using other table columns, showing competitive performance with improvements for longer sequences and through multi-task learning.
Missing data in tabular dataset is a common issue as the performance of downstream tasks usually depends on the completeness of the training dataset. Previous missing data imputation methods focus on numeric and categorical columns, but we propose a novel end-to-end approach called Table Transformers for Imputing Textual Attributes (TTITA) based on the transformer to impute unstructured textual columns using other columns in the table. We conduct extensive experiments on three datasets, and our approach shows competitive performance outperforming baseline models such as recurrent neural networks and Llama2. The performance improvement is more significant when the target sequence has a longer length. Additionally, we incorporate multi-task learning to simultaneously impute for heterogeneous columns, boosting the performance for text imputation. We also qualitatively compare with ChatGPT for realistic applications.