Mining of book-loan behavior based on coupling relationship analysis

Book-loan system is the most significant part among all the library services. In this study, we aim to provide a new perspective for how to improve the quality of library services based on the relationship between the characteristics and book-loan behavior of students. Most existing works regarding this issue are limited to statistical study on monolayer networks. Different from them, we focus on the characteristics of students, such as class, term, region and major. We first construct a two-layer book-loan network model to portray two components (i.e., students and books) and the relationship existing in the real book-loan system. Then, we integrate different data extracted from a real book-loan dataset by different rules into the model for analyzing the effect of different characteristics of students on their book-loan behaviors. The results show that the book-loan behavior of students is relevant to the major and term, but not to the class and region. We also find that the book layer of such network exhibits the scale-free and small-world properties.