The Research And Implementation Of Ontology Construction Method Based On The Big Data In Traffic Domain

With the rapid growth of China social economy and Internet technology and the constant depth of information construction in transportation area,the traffic field produces a mass of data,these data which have the characteristics of heterogeneous,massive and difficult to share are stored as different forms in different data centers.Although the Ministry of Transport has forced the implementation of the <traffic information essential data element> to unite data structures,there are still a lot of overlaps between these data which caused some problems such as repeated definitions of data,large amount of data and so on.Since ontology was introduced into the field of knowledge engineering,it has played an important role in knowledge structure,sharing and reusing.This paper proposes an ontology model based on the combination of ontology technology and data element,and studies the method of using the massive traffic data to construct ontology in transport area.The main contents include:Firstly,this paper discusses the structure and category of traffic information basic data element and explores the organization form and construction method of ontology,analyzes the characteristics and similarity between traffic information basic data element and ontology model.On this basis,the paper proposes the ontology hierarchy model in traffic domain based on text big data.Secondly,this paper studies the construction method of each layer in the domain ontology hierarchy model.In the first place,We construct the basic data element ontology of traffic information according to <traffic information essential data element>,which achieves the construction of common ontology layer based on data element.We put forward the whole framework to the traffic domain ontology hierarchy and present a detailed description for the subsidiary ontology framework in highway field.In the second place,we conduct the study and experiment about the Chinese text participle based on traffic dictionary and the term extraction based on Hidden Markov model and the concept relevance judgment to extract the concept of constructing ontology from traffic text big data.And then,we complete the construction of metadata layer through K-means clustering and manual cleaning.In the end,we treat data element ontology as a source in order to realize the construction of subdomain ontology layer,using the method of knowledge engineering to obtain concepts from metadata layer.Thirdly,we design and implement the construction of prototype system in traffic domain ontology,and verify the ontology model through comparison experiments.In this paper,the construction of the traffic domain ontology is semi-automated,it provides a feasible method to overcome the shortage of traffic information data element and construct domain ontology in accordance with the characteristics of transportation industry.