基于领域本体的游戏攻略文本标注算法研究与实现

摘要

Nowadays,game web portals set up plenty of websites,providing game information and related walkthroughs,for players to enhance their gaming experience.However,these sites has obvious isomerism and lacks unified hierarchy.Thus,an annotation algorithm based on domain ontology is proposed.It is started with a data fusion step from a set of web portals to build a game domain ontology.Meanwhile,the neologism discovering algorithm is optimized according to its application in game domain,and the semantic annotation for walkthrough text is further developed.Thus these semantic tags not only embody the intent of each guides,but also serve the semantic search for walkthrough text.Experiments have proved that the proposed ontology construction method is scalable.Moreover,the optimized domain vocabulary discovering algorithm has a better result compared with the traditional segmentation tools.%游戏门户网站为提升玩家们的游戏体验,建立了大量站点用以提供游戏资讯及相关攻略.然而这些站点间异构现象明显,且缺乏统一的知识体系.提出基于领域本体的文本标注算法,通过融合站点间的数据,构建游戏领域本体.同时,针对游戏领域的应用,优化了新词发现算法,并进一步对攻略文本进行语义标注.通过这些语义标签,不仅能直观地了解攻略中的内容,也能更好地为攻略文本的语义检索服务.实验证明,所提出的本体构建方法在游戏领域具有一定的推广性,同时游戏领域词汇发现算法与传统的分词工具相比也取得了更好的结果.