The Structure,Service Efficiency And Lending Behavioral Characteristics Of The Library Resources

From paper documents to electronic resources,the isolated one to widely interconnected and shared,the library is generally become a place of gathering big data.Library not only accumulated a large amount of books and their copies,but also the development of digital library created increasingly rich electronic resources,especially the interlibrary loan and document delivery integrate libraries dramatically.Resources integration with archives and museums further expand the boundary of the library.On the other hand,open access repositories and search engines also provide a large and convenient information resource,and play an increasingly important role in academic research.Moreover,the interactive data between library and readers(lending behavior record,retrieve information,web visit etc.)is also the typical big data(machine data),which enrich the types of library sources and provide the possibility of expansion of knowledge service.However,the value of big data is not storage,but usage;not single database,but connection of them.So this paper not only extracts lending behavioral records but also with questionnaire supplemented.By cross validation between different data sources,we analyze the structure and service efficiency of library resources,discuss the lending behavioral characteristics,study the main influence factors of library resource utilization,and put forward practical suggestions.This paper uses the sample of graduate student with high performance of utilization.We sample 600 graduate students of the Dongbei University of Finance and Economics,employ three-stage PPS sampling,and acquire 526 valid questionnaires with 87.7%response rate.Lending behavioral records extract one-tenth of overall graduate students,which constitute the sample size 262.In the process of investigation,we adopt the following ways to ensure high precision:1.Make comprehensive survey plan in detail and enforcestrictly;2.Execute pre-research to avoid investigation risk;3.Prove the consistency of content and effectiveness in the questionnaire design through the reliability and validity tests;4.Employ three-stage PPS samplings guarantee randomness.In the way of data-collection,we consider the "hear more sides to make decision",and integrate the questionnaire survey with lending behavioral records to obtain sample data.Step on data cleaning,we not only consider the combination problem of multiple choice questions,but also convert nominal data into ordinal data and transform lending behavioral records to data in basket type suitable for mining analysis of user behavior.In data analysis phase,we employ variance analysis,contingency analysis,nonparametric test and other basic statistical analysis methods,and combine generalized linear mixed model and association rules to evaluate the statistical characteristics of library resource utilizationThe following conclusions are obtained:First of all,students prefer books rather than electronic resources and open resources such as search engine.A traditional explanation considers the cause of reading habit,however,according to the survey that the large amount of borrowing of literature books largely contribute to the paper books circulation.In addition,the absolute number of paper books circulation is still dominant,but according to lending behavioral records,lending rate reveals the downward trend from 2005-2014.The use efficiency after borrowing is relatively high,only 2.7%of the readers have never read.The third,the peak of lending in autumn semester(November)is significantly later than the spring semester(March).The role of the library plays we find that"make friends" and "consult classmates" exist significant differences in gender.The main purpose to go to the library is not "borrow books and get information"(29.66%),but "self-study"(64.64%),the proportion of "attend lecture or activities"is less than 1%.At last,lending behavior seemingly follows "Pareto principle" that the top three categories sorted in the lending rate support for above 70%library circulation,while most categories are not in their favor.To improve circulation efficiency and expand the reading scope,we use association rules to explore the correlation between categories.