A study on how to construct the prediction model of library lending of university library

As for evaluating an administrator of a library, library lending is an important indicator. Furthermore, if library lending can be known previously, then the information may help an administrator make the annual plan. The change of library lending depends on the change of the number of readers and on the change of the books that a library possesses. Library lending has a strong association with the number of readers. The author has investigated library lending and the number of readers of a university library in China for 18 school years. It is found by the author that the sample data of library lending and the number of readers fit the simple linear regression model. Two methods of test, such as coefficient of simple determination and t-test prove that a strong linear relationship exists between library lending and the number of readers. The result of t-test is acquired by running the function of regression of data analysis of Microsoft Excel. Then Based on the theory of the simple linear regression analysis, the prediction equation is obtained. The author gives the methods of calculating confidence intervals of the mean library lending and prediction intervals of a single library lending in the coming school year.