所属单位：经济与管理学院
发表刊物：Lect. Notes Comput. Sci.
摘要：This paper aims to solve the problem of identifying commodity names in the field of cross-border e-commerce by using some machine learning algorithms. As far as we know, this is the first attempt to use machine learning algorithms to solve this kind of problems in this field. A model of commodity name recognition algorithm based on the SVM and TF-IDF models is proposed. For 115,521 commodity description texts containing 2,128 different commodities, experiments show that the recognition accuracy of the algorithm is 91% and the recall rate is 93%. At the same time, compared with the traditional manual method, the algorithm can improve the efficiency of commodity export declaration by about 20% during the "Double Eleven" period. It is proved by experiment and practice that this algorithm is reasonable, effective and practical. © 2019, Springer Nature Switzerland AG.
ISSN号：0302-9743
是否译文：否
发表时间：2019-01-01
合写作者：Li, Xiaofeng,Li, Chi,,Guo, Xiaoyu
通讯作者：马静
发表时间：2019-01-01