Research On Scientific Data User Relevance Clues And Criteria And Relationships

Scientific data resources are the foundation of scientific research.With the development of science and technology,the storage scale of data resources is increasing,but the low efficiency of obtaining scientific data is becoming more and more prominent.The direct cause of this phenomenon is lack of efficient search tools,and the fundamental cause is the lack of understanding of the principle and mechanism how users determine the scientific data relevant.Especially,the understanding of basic concept of scientific data user relevance clues and criteria,and the relationship between them are weak.In this paper,we take the scientific data as the target information type,and carry out the empirical research on the relationship between the scientific data user relevance clue and criteria.It aims at deepening the understanding of scientific data user relevance judgment mechanism,and exploring the information processing of human brain on data clues and criteria,so as to facilitate computer simulation and provide an algorithm and theoretical foundation for developing intelligent search engines.In this study,the research was conducted with the methods of literature survey,constitutional experiment,questionnaire survey,interview,thinking aloud and grounded theory.The main research contents are as follows:(1)The study of scientific data the user relevance clues.Through observing and interpreting the subjects' relevance judgment behavior,the relationship between the characteristics of scientific data and relevance judgments in user long memory was explored,and clues set was set up.(2)The study of scientific data the user relevance criteria.Through comparative study,the relationship between different target information type differences and relevance criteria differences was explored,the connotation and classification of scientific data relevance criteria were further understood and modified,and a scientific data relevance criteria set was established.(3)The relationship between scientific data user relevance clues and criteria.Based on the clues set and criteria set in recognition,study the stimulus and response behavior in the process of relevance judgment,pay attention to the explanation of the thinking course,and establish the relationship between scientific data clues and criteria on the basis of data statistical analysis.By studying the behavior of scientific data user relevance judgment,explore scientific data user relevance clues,criteria and the relationship between them,in order to better understanding of the scientific data user relevance judgment mechanism and principle,providing theory and algorithm foundation for intelligent search engine design and develop.The main achievements of the research are as follows:(1)Scientific data clues can be divided into 5 categories: content clues,quality clues,data acquisition clues,external evaluation clues and professional clues.When the user's professional field changes,the professional clues have changed greatly.(2)There are 12 criteria for user relevance judgment,which can be classified into 2 categories: Data ontology and data availability.Data ontology is user evaluation criteria for data physical entities,including topic,availability,normalization,quality,authority,timeliness and novelty.Data availability is the criteria for user to evaluate data to be used,including professional requirements,comprehensibility,availability,convenience and comprehensiveness.Ontology and availability play a role at the same time in the user relevance judgment.But if we cannot get enough information or want to search quickly,only evaluate the ontology of data can also meet the requirements.The data availability evaluation alone can not get the judgment result.(3)The relationship between clues and criteria can be divided into 3 categories: one stimulus multiple responses,multiple stimuli one response and multiple stimuli multiple responses.Regression analysis found that there was a positive correlation between clues and criteria,but the existing clues had lower standard regression coefficient.In data relevance judgment,users need to evaluate multiple clues in a comprehensive way.The existing clue system can basically meet the user's retrieval requirements,but there is still a certain gap to meet the requirements of users' precision and personalization.The optimization design of data clues and the rational use of professional clues will help to improve the efficiency of data retrieval.The innovation points of this thesis embodied in the following aspects:(1)The conception of relevance criteria is divided into two concepts,relevance clues and criteria.The relationship is explained as the relationship between stimulus and response,which leading to a better people's understanding of the scientific data user relevance determining mechanism.(2)Clues set and criteria set which affect scientific data user relevance judgment are proposed,and the relationship between clues and criteria is established.This research provides the user level basis for the scientific data retrieval system analysis and design and the improvement of the system performance.