Research On Address Matching Algorithm Of Place Name

Address is a kind of dispensable basic geographical information and social public information which is used to describe a geographical entity.The traditional data processing of address is usually by the means of some acritical approaches like the national geographical survey,the field investigation and so on.However,this kind of way may lead to high workload,thus causing too much expense of manpower with low working efficiency and timeliness,in return.With the development of Internet and the coming of big data era,the knowledge system of this subject is continually upgrading.Therefore,attaining the geographical data information from the abundant information resource on the Internet becomes a new way to obtain and renew the database of address.Owing to the lack of a unified standard of naming and the diversified and complicated forms of data on the Internet,these drawbacks will lead to the deficiency of effective information,which may give many misunderstanding to obtain and renew the database of address.In order to extract the information from the disorganized data of the Internet,we should learn from the technology of Chinese Natural Language Processing,Information Extraction,and Machine Learning,ML.Combining with the technology of Natural Language Processing and Web Spider,it will deal with some typical problems that we confront with in traditional data processing of address.The data of address is just a series of strings describing the geographical location.We need to use the geographical name address matching technology to map the address information input by the user to the accurate spatial position coordinates of a geographical entity in the database.The geographical name address matching technology is an important basis for space navigation services such as modern traffic navigation.In summary,the main contents of this paper are as follows:(1)This paper studies the data form of standard place names and addresses,and the difference between the data of network place names and the data of standard place names and addresses,so as to provide the precondition for the following data storage of place names and addresses crawled in the network.(2)Combining network crawler and Chinese natural language processing,we can get the relevant data of place names and addresses from the massive data,which can solve the problems of high labor cost,large workload,low efficiency and low timeliness of using traditional surveying and mapping methods to obtain place names and addresses information data.(3)By using Chinese natural language processing technology,the user's input intention can be reasonably inferred,and the input errors or missing address elements can be solved,thus the accuracy and intelligence of place name and address matching can be greatly improved.