The Analysis of interlibrary Loan Transactions in Birjand University of Medical Sciences using Data Mining Algorithms

Data mining is a process for discovering meaningful relationships and patterns from data. Identify behavior patterns of libraries users can helps improve decision-making in libraries. This study aimed to analyze the interlibrary loan transactions in Birjand University of Medical Sciences using data mining algorithms. In this descriptive study, knowledge discovery and data mining were used to determine patterns of interlibrary transactional and circulation data in Birjand University of Medical Sciences. Information on the members and circulation transactions was extracted from the library database, and saved in Excel file format after adding some computational fields. Data were processed using Microsoft SQL Server 2008 Data Mining Add-Ins, Office 2007. Classification Matrix was used to evaluate the accuracy of the models. The findings were reported in tables and graphs. 394,011 records of library circulation transactions from 5600 members were collected up to 20th of June 2013. Flow of transactions per semester was of a regular sequence. About 87% of the borrowed books were returned less than 46 days and 96% of students fail to return books on due date and the delay time is less than 18 days. The members borrowed 35 books on average during their membership. Related subject of interest to members was discovered through association rule algorithm. Data mining can be used for identifying behavior patterns of members, classifying and identifying factors affecting the duration of book loan and delay. Revealed patterns are suggested for decision-making in human resources management and development of regulations on library circulation services.