A Disambiguation Archiving and Storage Method for Data of Scientific Research Achievements

Abstract

The present invention provides a scientific research achievement data disambiguation archiving storage method, including: S1, processing and completing the scientific research achievement data that has been archived and to be archived, and saving it to the database; S2, calculating the similarity of collaborators, and filing if they match , if not, enter S3; S3, cluster the data that has been archived and the data to be archived; S4, calculate the distance from the center point of each cluster of the data to be archived to the center point of each cluster of each scientific researcher with the same name, and obtain the closest distance Scientists who belong to the cluster; S5. Establish a feature relationship graph; S6. Calculate the similarity probability between the data node to be archived and each data node that has been archived, and calculate the average and variance, and compare it with the threshold to complete the archive. The solution proposed by the present invention does not require data labeling training, and is more practical in most scientific research personnel systems, and can quickly realize data disambiguation and simultaneously improve the accuracy of disambiguation.

Description

technical field [0001] The invention relates to the field of data processing, in particular to a method for archiving and storing scientific research achievement data disambiguation. Background technique [0002] With the popularization of the Internet, different institutions or departments currently have their own online scientific research personnel systems or scientific research personnel information databases, and a large number of scientific research personnel achievement information will be added from time to time. These data must be accurately archived in the system. Authors already exist. Archives, the problem of researchers with the same name is an urgent problem to be solved by this type of system. [0003] Existing disambiguation methods of the same name almost all transform the problem into a related clustering or classification problem for machine learning. In the process of data incremental disambiguation, most scholars currently use the supervised disambiguat...

AI Technical Summary

Problems solved by technology

This method has an obvious shortcoming. The binary classification method does not consider the global distribution characteristics of the author's literature. If a researcher's scientific research data research direction is completely different from other literature, it will lead to misclassification.

[0004] Usually, the disambiguation method based on supervised learning is better than other methods, but in practical applications, it is impractical and expensive to manually label the data of large-scale scientific research results database

At present, most scientific researchers disambiguate with the same name, and disambiguation is done based on the paper data set, without considering the data characteristics of other scientific research results, it will appear that a researcher's paper is completely different from other documents, but may be related to a certain patent, or Other types of scientific research data belong to the scientific research achievements of the same period

Images