Natural language processing in “Bullet Screen” application

The main purpose of nature language processing is to process, understand, and apply all kinds of human beings' languages in written or oral forms. In recent years, “Bullet Screen”, which allows viewers to post bullet-like, real-time comments on screen during their watching films, is an emerging craze in online video sites, especially in China and Japan, mainly popular among young people for their social interactivities. Meanwhile, since “Bullet Screen” can be regarded as a novel type of natural language, processing, restoring, and organizing them play an important role in providing retrieval platforms for users so as to help them find out highlights from oceans of videos. Taking Bilibili, a Chinese video website, as example, this paper comes up with a new algorithm called Integrated Sort for “Bullet Screen” (ISB) base on BM25 in order to put “Bullet Screen” processing into practice. Empirical study shows that ISB based on BM25 can anatomize “Bullet Screen” language and mine more details about audio as well as video information. Besides, an interactive platform about “Bullet Screen” could enormously refine viewers' lives of entertainments.