The smallest extraction problem

We introduce landmark grammars , a new family of context-free aimed at describing the HTML source code pages published by large and templated websites therefore effectively tackling Web data extraction problems. Indeed, they address inherent ambiguity HTML, one main challenges extraction, which, despite over twenty years research, has been largely neglected approaches presented in literature. then formalize Smallest Extraction Problem (SEP), an optimization problem for finding grammar that best describes set contextually extract their data. Finally, we present unsupervised learning algorithm to induce from sharing common template, automatic system. The experiments on consolidated benchmarks show approach can substantially contribute improve state-of-the-art.