Web Content Extraction - a Meta-Analysis of its Past and Thoughts on its Future

TL;DR

This paper provides a meta-analysis of web content extraction algorithms, revealing their limitations in handling modern, evolving web pages and discussing future directions to improve robustness and adaptability.

Contribution

It offers a comprehensive review of existing extraction methods, highlighting their weaknesses and proposing future research directions to address these challenges.

Findings

Most extractors do not handle modern web pages well.

Wrapper induction methods break as Web changes.

Heuristic extractors also degrade over time.

Abstract

In this paper, we present a meta-analysis of several Web content extraction algorithms, and make recommendations for the future of content extraction on the Web. First, we find that nearly all Web content extractors do not consider a very large, and growing, portion of modern Web pages. Second, it is well understood that wrapper induction extractors tend to break as the Web changes; heuristic/feature engineering extractors were thought to be immune to a Web site's evolution, but we find that this is not the case: heuristic content extractor performance also tends to degrade over time due to the evolution of Web site forms and practices. We conclude with recommendations for future work that address these and other findings.