Survey on Semantic Table Interpretation

Tracking #: 1946-3159

This paper is currently under review
Authors: 
Lahiru de Alwis
Achala Dissanayake
Manujith Pallewatte
Kalana Silva
Uthayasanker Thayasivam

Responsible editor: 
Jens Lehmann

Submission type: 
Survey Article
Abstract: 
The web contains a vast amount of tables which provide useful information across multiple domains. Interpreting these tables contribute to a wide range of Semantic Web applications. Aligning web tables against an ontology to understand their semantics is known as Semantic Table Interpretation (STI). This paper presents a survey on Semantic Table Interpretation(STI). Goal of this paper is to provide an overview of STI algorithms, data-sets used, and their evaluation strategies and critically evaluate prior approaches. In the effort of providing the overview we developed a generic framework to analyze STI algorithms. Using this framework we analyzed the existing algorithms and point out their strengths and weakness. Additionally this enables us to categorize the prior works and be able to point out the key attributes of each categories. Our analysis reveals that search based approaches are better in terms of accuracy and overall completeness, while other categories perform better only in annotating columns with high precision. Also, We present the evaluation methodology utilized in algorithms and discuss the limitations of it while providing suggestions for future improvements. In addition, we point out the design choices in building an STI and their associated trade-offs, which could be of value for the future STI algorithm developers and users.
Full PDF Version: 
Tags: 
Under Review