科研之友
微信
新浪微博
Facebook
分享链接
摘要
Obtaining a useful complete overview of Web-based product information has become difficult nowadays due to the ever-growing amount of information available on online shops. Findings from previous studies suggest that better search capabilities, such as the exploitation of annotated data, are needed to keep online shopping transparent for the user. Annotations can, for example, help present information from multiple sources in a uniform manner. In order to support the product data integration process, we propose an algorithm that can autonomously map heterogeneous product taxonomies from different online shops. The proposed approach uses word sense disambiguation techniques, approximate lexical matching, and a mechanism that deals with composite categories. Our algorithm's performance compared favorably against two other state-of-the-art taxonomy mapping algorithms on three real-life datasets. The results show that the F-1-measure for our algorithm is on average 60% higher than a state-of-the-art product taxonomy mapping algorithm.
- 出版日期2016-3-1
