Product Information Words Recognition Based On Conditional Random Field In Electronic Commerce

With the rapid evolution of the Internet technology, the electronic commerce which is based on the virtual economy has become an indispensable part in one’s life. Meanwhile the information brought by the electronic commerce is expanded, it becomes larger in which how to find the product needed accurately through the search engine has become a big problem. With the introduction of Chinese word segmentation technology into the electronic commerce, we can make the product description information segmentation more accurately to improve the retrieval precision of search engines.In this paper, we first apply the Conditional Random Field (CRF) to the study of the product description information segmentation in the electronic commerce field. On the basis of combined with the feature of the electronic commerce, we select some product description information from a shopping website as the experimental corpus and use the experimental tool CRF++ to validate the experiments of Chinese word segmentation and out-of-vocabulary word recognition. Furthermore, this paper conducts various experiments to compare segmen-tation performance under different situations, such as different models, different feature tem-plates. Experimental results and analysis provide that CRF model shows very good perfor-mance.