Research On Intelligent Query Technology Of Ship Information Based On Machine Translation And Knowledge Graph

Querying ship information is of vital importance to the entire shipping industry chain,and is also the most basic function of many ship service platforms.However,the traditional ship information query platform often just mechanically compares the matching relationship between query words and database data,and then lists the query results directly to the user,but does not intelligently understand the user's true intentions,providing more accurate and more relevant Sexual knowledge.In response to this problem,this article has deeply researched the intelligent query technology of ship information,using the latest deep learning and machine translation methods to build a translation model of AIS Pinyin to Chinese,which is used to solve the ship's multiple matching methods,Port and route data to build a Chinese knowledge map of ship information,and finally realize intelligent query of ship information.In summary,the work of this article mainly includes:1)Analyze the main problems of the intelligent query service of ship information,and combine the current research status of machine translation and knowledge mapping technology at home and abroad to propose a feasible technical solution;2)Construct a corpus of standardized ship information,including a corpus corresponding to the AIS pinyin ship name and Chinese ship name,and a knowledge graph corpus of related information such as ships and ports,routes,and trade areas3)Construct an intelligent ship name translation model based on the combination of Transformer and Hidden Markov Chain(HMM),which realizes the accurate translation of AIS pinyin ship names into Chinese ship names,and solves the intelligent matching of AIS information with other ship information.Trained and tested on the corpus data set based on the ship database of Shipun.com,the final accuracy rate of the model reached 98.92%,and the accuracy rate on the same data set was significantly better than the ship name translation model based on SMT and Seq2Seq;4)According to the attribute association between ship-related information,the information of ships,ports,routes,operating companies,trade areas and countries are materialized and connected to form a knowledge graph,and then a ship information query model is constructed to realize intelligent query of ship information,Solve the problem of accurate answer and efficient multi-modal search.In summary,this article uses the latest deep learning and machine translation methods to study and build an intelligent translation model of AIS pinyin ship names to Chinese character ship names,and further uses knowledge maps to study and build intelligent query technology for ship information It is a useful attempt to apply artificial intelligence in the field of shipping.It not only has theoretical significance,but also has certain practical value for improving the service level of shipping information.