Extraction of Product Information Object for Trustworthiness

摘要

Online market brings a wealth of information to consumers. However, online market is also flooded with fake and shoddy products. Trustworthiness is a solution to solve these problems which is raised by the authors' laboratory. The model of trustworthiness evaluates the product in three dimensions, which is authenticity, reliability and credibility. But the data to analyse of trustworthiness evaluation is mainly from manual import. Product Information Object is the organic whole of product information and product reviews. Extraction of Product Information Object for Trustworthiness in this paper will expand the data source of product information and product reviews for Trustworthiness effectively. In this paper, combining with project background, we made a deep research on pages capture, pre-treatment of web page, extraction of web information. And we find out that the accurate rate of extraction of product information is not high and the coverage rate of extraction of product reviews is very low. To solve these problems, we improve the related algorithms, and design a module which can extract product information objects automatically. In this way, the requirement of analysis data for Trustworthiness will be satisfied.