Learning Parliamentary Profiles for Recommendation Tasks
CAEPIA(2015)
摘要
We consider the problem of building a content-based recommender system in a parliamentary context, which may be used for two different but related tasks. First, we consider a filtering task where, given a new document to be recommended, the system can decide those Members of the Parliament who should receive it. Second, we also consider a recommendation task where, given a request from a citizen, the system should present information on those deputies that are more involved in the topics of the request. To build the system we collected, for each Member of the Parliament, the text of corresponding speeches within the parliament debates and generated, with different techniques, a profile that was used to match against the input document or request. We tested our methods using the documents of the regional Andalusian Parliament at Spain, obtaining promising results.
更多查看译文
关键词
User profiles,Content-based recommender systems,Information filtering,Information retrieval,Parliamentary documents
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络