Research On Attribute Identity And Retrieval Method Of Open Government Data

With the advent of the era of big data,data becomes a new production factor,and people gradually realize the value of data.To satisfy the social requirement of data,government data resources have gradually been opened to public,and people could find,download,analyze and use them through open data platform to achieve the value of data.However,there are many obstacles in this process currently,showing some serious ploblem such as "digital divide","poor quality data",and "data not easy to find",which greatly affected the social development and utilization of data.To solve those problems,this paper presents an organization framework of open government data from the overall level,and conduct in-depth study on the key issues.The main contents are as follows:(1)Propose an organizational framework for open government dataTo realize interoperability of open data,user participation,and the value of data,we design the organizational framework for open government data,which involves three processes: data description and publishment,data link and retrieval,and data consumption and reuse.Attributes identification and data retrieval are two key issues,which can achieve interoperability and precise location of the data,and improve the quantity and quality of using data.Then illustrates the relationship between attribute identification and data retrieval,that is,attribute identification can help for data retrieve.(2)Identify same properties for open government dataAccording to content similarities and same context of same attributes,this paper proposes an integrated method based on attribute values and context.The main process is follows,calculate the similarity between attributes based on attribute values and context,and calculate integrated similarity.Then,according to the results of top k attribute similarity,human judge and evaluate the accuracy of the method.The experimental results show the integrated method can improve accuracy of using a single method to identify attribute.Therefore,prove the method is feasible.(3)Retrieval method research for open government dataDesign retrieval system for open government data,of which the relevance ranking and retrieval are the key parts.For many shortcomings of naive ranking,propose the method of feature ranking(i),that is,design features for open data retrieval,and train ranking function using artificial labeled sample.Consider heterogeneous properties have some impact on ranking results,so apply attributes identification to rank model,adding the feature of whether hits same attributes,namely feature ranking(ii).To achieve the algorithms,establish the inverted index,where offset is represented by two-dimensional coordinates,compared with traditional method.The experimental results show the feature rank(ii)model is feasible.