We assimilate the extracted information from a conference website to acquire the clean and high superiority academic data. This research has subsequent contributors: We propose a novel vision-based page segmentation algorithm, which use DOM tree to compensate the information loss of classical vision-based segmentation algorithm. We transform the conference Web material extraction which is difficult into a classification problematic, and categorize text blocks as predefined sets permitting to vision, key disputes, text and content information. We improve the classification quality by post-processing. Our experimental results on real-world datasets shows that our method is highly effective and efficient for extracting academic information from conference pages.

Design Implementation