We develop a network reconstruction model based on entropy maximization considering the sparsity of networks. We reconstruct the interbank network in Japan from financial data in individual banks' balance sheets using the developed reconstruction model from 2000 to 2016. The observed sparsity of the interbank network is successfully reproduced. We examine the characteristics of the reconstructed interbank network by calculating important network attributes. We obtain the following characteristics, which are consistent with the previously known stylized facts. Although we do not introduce the mechanism to generate the core and peripheral structure, we impose the constraints to consider the sparsity that is no transactions within the same bank category except for major commercial banks, the core and peripheral structure has spontaneously emerged. We identify major nodes in each community using the value of PageRank and degree to examine the changing role of each bank category. The observed changing role of banks is considered a result of the quantitative and qualitative monetary easing policy started by the Bank of Japan in April 2013.