HybEx: A Hybrid Tool for Template Extraction
Julián Alarte, Josep Silva · Companion Proceedings of the Web Conference 2022 · 2022
HybEx is a site-level web template extractor that combines two algorithms for template and content extraction: (i) TemEx, a site-level template detection technique, and (ii) Page-level ConEx, a content extraction technique. The key idea is to add a preprocess to TemEx that removes the main content inferred by Page-level ConEx. It is a fact that adding this new phase to the TemEx algorithm involves an increase of its runtime, however, this increase is very small compared to the TemEx runtime because Page-level ConEx is a page-level technique. On the other hand, HybEx improves the precision, recall, and F1 of TemEx. This paper describes the new template extractor and its internal architecture. Furthermore, the paper also presents the results of its empirical evaluation.