A WebExtension framework for experimentation and evaluation of webpage segmentation methods

초록

Current webpages contain areas with different functions and contents. Many studies and applications have used webpage segmentation methods to separate these areas or extract only specific areas for their purposes. Examining these methods requires laborious tasks, such as collecting many webpages, inspecting them with human participants, and applying various performance metrics to their results. Therefore, we developed a WebExtension (browser extension) framework to support the examination and analysis of webpage segmentation methods. This framework can build a WebExtension to collect webpages, curate data for labeling web documents, evaluate methods, and measure the results with various performance metrics in a web browser environment. Furthermore, researchers can use preloaded well-known methods and metrics in the framework and add more methods and metrics for their research purposes.

키워드