%0 Journal Article %T The application of multivariate analysis approaches to designing NSBM model considering undesirable variable and shared resources %J Scientia Iranica %I Sharif University of Technology %Z 1026-3098 %A Ghasemi Toudeshki, H. %A Najafi, E. %A Hosseinzadeh Lotfi, F. %A Movahedi-Sobhani, F. %D 2020 %\ 04/01/2020 %V 27 %N 2 %P 893-917 %! The application of multivariate analysis approaches to designing NSBM model considering undesirable variable and shared resources %K confirmatory factor analysis %K structural equation %K network data envelopment analysis %K network slacks-based measure %K universal banking system %R 10.24200/sci.2018.5578.1392 %X Due to the competitiveness of banking industry and increasing bargaining power of customers, evaluation of the banks’ performance is crucial to better serve the classified customers in a universal system .In this paper, with consideration of segmenting the customers into personal and business ones, methods such as confirmatory factor analysis (CFA) and structural equation model (SEM) have been used in selecting appropriate variables of the network data envelopment analysis (NDEA) model based on network slacks-based measure and consideration of the undesirable variables and shared resources. The SEM model has been used to establish a proper connection between the different dimension of the NDEA model and CFA model has been used to identify the importance of each dimension. Also, the proposed model has been used to measure the Operational and decomposed universal efficiency of one of the Iranian bank branches (Day Bank). The results show that the extracted model provides managers with a suitable perspective in adopting appropriate policies to promote their performance in the different sectors, including deposit attraction, financial serving personal and business banking customers, and profit generation, and also in comparing them in the different dimensions of the model. %U https://scientiairanica.sharif.edu/article_21053_09c3f73f1dc76477e980df5afeb9bac9.pdf