基于内容和背景的微博问答问题推荐

Question recommendation of microblog QA based on content and background

新浪微博的新功能微博问答一经上线就抢占了大部分内容付费的市场,如何对特定的博主提出其愿意回答并且围观量多的问题就成了我们关注的重点。针对问答关键词的推荐问题,提出了基于AW-LDA模型的用户关键词挖掘方法,并结合微博内容和背景并存的特点,采用了基于内容和背景的用户相似度分析方法。通过进行对比试验,结果表明:该用户关键词挖掘方法和问题关键词推荐方法相较于传统方法推荐的问答关键词推荐效率提高了9.15%,推荐的关键词收益率提高了15.53%。

Microblog QA has seized most of the market of content payment not long after its on line as a new feature of Sina microblog,so it turns out to be a problem we focus on that how to make the problem we ask tending to be answered and profitable. Aimed at the question of mining the QA keywords,this paper proposed a method of mining user keywords based on AW-LDA model, and taking the characteristics of the content and background of microblog into consideration,this paper raised a method of user similarity analysis based on content and background. Contrast test shows that,compared with the traditional methods,the keyword recommendation efficiency has been increased by 9.15%,and the yield of the recommended keyword has increased by 15.53%.