The Books In The Library's Network Topology Analysis

As the general abstract and description way of complex system, complex networksresearch made in-depth development in recent years. Domestic and foreign scholarshave done a lot of empirical researches on complex systems from the network structure,dynamics on the network, network evolving, et al, and have made a series of theoreticalresearch results. Using complex networks theory to study complex system can deepenpeople’s further understanding of the system structure.By collecting the one-year data of book-borrowing in library of University ofShanghai for Science and Technology, this paper research on its topological structureusing complex network idea and theory. The former research most focus on topologicalanalysis of un-weighted networks which can partly reveal topological character ofnetworks but lack in deep understanding network function. Based on this, this articlestudies the library network by three weighted methods. In the cooperation andcompetition network, we introduce the sum reading time that a reader borrows a book(with the same call number) as the weighted. Compared with the un-weighted network,the weighted network can not only describe the cooperation-competition situation andconfiguration, but also the competition results accurately. In the single-mode networkprojected from the readers to the books, we use the number of the repeated edges as theweighted value. In the library personal recommendation system，we adopt a newproper weighting method. Through the historical borrowed books for a reader, we willgive some suggestions to him when he needs some books. No matter what types ofweighted method, we find that distributions of act size, act degree, node strength anddegree of single-node networks all decay exponentially and also prove single-nodenetworks having characteristics of small-world. There is teacher-student cluster,specialty cluster or interests cluster in the action of borrowing and the degree of clusteris very high.Secondly, through the research of the relevant time statistic of library lendingmode, we find the statistical characteristics of a group users, that is, the distribution ofthe lending quantity of time period in every working day is similar. Then, through thestudy of the interval time between the borrowed and returned time, we find that thereading time distribution shows an approximately power-law decaying, changing at thetime limit, which provide more empirical foundation to a new discipline named human dynamics. Finally, we bring forward the deficiency of this paper and discuss the fieldsabout farther study in book-lending relations network in library.