Selecting strategic partner for tax information systems based on weight learning with belief structures

Abstract

To select strategic partners for local tax departments to provide tax information systems (TISs) and long-term and quality services, which is generally modeled as a multiple criteria decision making (MCDM) problem, this paper proposes a new MCDM method with belief structures based on weight learning. To facilitate modeling and analyzing this problem, a criterion framework for selecting TIS partners is constructed by considering existing criteria for selecting strategic partners and the special requirements of TIS. Within the developed criterion framework, two ways are designed by using the maximum likelihood estimation and the Hurwicz rule to learn criterion weights from the overall comparisons between specific partners and the individual assessments of the partners on each criterion that are denoted by belief structures. One way utilizes the precise optimism degree in the Hurwicz rule and the other utilizes the interval-valued optimism degree. The minimum and maximum expected utilities of each partner are determined by using the learned weights of criteria to compare different partners. The problem of selecting an appropriate TIS partner for a local tax department located in Ma anshan, Anhui Province, China, is analyzed to demonstrate the proposed method.