Design And Implementation Of Vertical Search Engine For Pharmaceutical E-commerce

With the implementation of "Internet +" strategic layout,e-commerce of our country has brought new opportunities for development in the pharmaceutical sector.Compared to traditional offline retail marketing methods,e-commerce platforms can help people quickly find suitable drugs and achieve maximum user privacy protection.However,the increase of similar drugs has caused serious information overload problem.Therefore,how to provide more accurate search results for user has become one of the key factors for the pharmaceutical e-commerce platform to increase users'purchase conversion rate.The traditional search approach based on keyword fuzzy matching in most medical e-commerce platforms has seriously affected the query accuracy and efficiency.It is of great practical significance to build an intelligent vertical search engine for the field of pharmaceutical e-commerce.According to the characteristics of the medical field,this thesis utilizes natural language processing,knowledge mapping,user behavior analysis,and machine learning technology to design several important components,which are used to build an intelligent vertical search engine for the field of pharmaceutical e-commerce.These constructed components include:the Chinese segmentation component in medical field based on HMM and Viterbi,the product relevance calculation component based on medical knowledge,the user's interest degree calculation component based on user visit and click behavior,and the comprehensive score ranking component based on multiple linear regression.By equipping with these components,our developed intelligent vertical search engine is able to fully mining the user's actual demand and thorough comprehend relevance of pharmaceutical products and user interests.As a result,the search efficiency and accuracy are improved significantly.The innovations of this thesis can be summarized as follows.(1)We design a vertical search engine based on medical knowledge and user preference for the field of medical e-commerce,which effectively solves the information overload problem and insufficient comprehensions of actual user demand problem;(2)We proposing a method based on the comprehensive ranking of medical knowledge maps and user interest preferences to assist users in rational drug use as well as takes into account the user's preference for the product;(3)We develop a vertical search engine prototype for medical e-commerce and compare it with certain company e-commerce platform search engine.The extensive test results show that the vertical search engine constructed in this paper can accurately understand the user's search semantics and accurately return the product search results.It has a higher precision and recall rate.