基于社会信息重排序的图书检索方法

发布时间：2019-08-26 13:21

【摘要】：随着社交网络的兴起和发展,互联网上出现了大量与商品有关的社会信息。如何利用这些社会信息结合商品元数据进行检索和推荐是信息检索领域中一个热门的研究问题。本文以社会图书检索为例,提出了一种通用的信息检索方法来解决这一问题。首先,通过分析原始图书数据集和图书的用户标签、用户评分和流行度等社会信息,从图书中提取不同的社会特征构建特征矩阵;然后分别计算图书在各种社会特征上的相似度,并使用不同的策略对搜索引擎返回的排序结果进行重排序;最后使用学习排序的方法进行重排结果融合,得到最终的图书检索结果。在实验中,使用该检索方法在INEX Social Book Search 2015和2016数据集上分别进行了训练和测试。结果表明,相比现有的技术,该检索方法能够有效提升图书检索的效果。
[Abstract]:With the rise and development of social networks, a large number of social information related to goods has appeared on the Internet. How to use these social information combined with commodity metadata to retrieve and recommend is a hot research problem in the field of information retrieval. Taking social book retrieval as an example, this paper puts forward a general information retrieval method to solve this problem. Firstly, by analyzing the original book data set and the social information such as user label, user score and popularity, different social characteristics are extracted from the book to construct the feature matrix, and then the similarity of the books in various social characteristics is calculated respectively, and the ranking results returned by the search engine are reordered by using different strategies. Finally, the method of learning sorting is used to merge the rearrangement results, and the final book retrieval results are obtained. In the experiment, the retrieval method is used to train and test on INEX Social Book Search 2015 and 2016 datasets, respectively. The results show that compared with the existing technology, this retrieval method can effectively improve the effect of book retrieval.
【作者单位】：北京科技大学计算机科学与技术系;
【基金】：国家自然科学基金“结合前馈和反馈机制的自然场景文本识别技术”(61473036)资助
【分类号】：TP391.3

【相似文献】