Research On Data Asset Quality Evaluation Methods For Commercial Banks To Eliminate The Preference Of Decision Makers Under Uncertain Environment

With the advent of the era of big data,the concept of data asset has gradually formed.Data asset quality is not only the embodiment of data asset value,but also the basis of data applications.Nowadays,while the speed of product and service innovation in the banking industry is increasing,the competition among commercial banks is intensifying.Big data application of data asset is the embodiment of the core competitiveness of commercial banks.Because data asset quality evaluation indicators are all qualitative indicators,as well as the existence of subjective confidence and preferences of decision makers in the actual evaluation process,coupled with the complexity of the evaluation environment,data asset quality evaluation is under uncertain environment.Therefore,uncertainty theory can be used to analyze and deal with the problem of commercial banks' data asset quality evaluation.The neutrosophic soft set is a good theoretical tool for describing uncertain information.It is also convenient to use linguistic terms such as “poor” and “general” to describe relevant indicators such as value density and safety factor in practical evaluation.Therefore,this research focuses on the commercial banks' data asset quality evaluation under uncertain environment,and constructs the data asset quality evaluating methods for commercial banks that eliminate the preferences of decision makers,by focusing on two forms of evaluation information,namely,neutrosophic information and linguistic information,respectively.This research aims to provide scientific guidance for the problem of commercial banks' data asset quality evaluation under uncertain environment.The main researches are as follows:(1)For the research on commercial banks' data asset quality evaluation with neutrosophic information,a new cosine similarity measure under neutrosophic soft environment is proposed by analyzing the defects of existing neutrosophic soft sets correlation cosine similarity measures.After that,the concepts of objective degree and credibility degree reflecting the subjective preference of decision makers are proposed,and the subjective evaluation information is fused by the Dempster-Shafer evidence theory,so as to construct a commercial banks' data asset quality evaluation method eliminating the preference of decision makers based on the neutrosophic soft sets.(2)For the research on commercial banks' data asset quality evaluation with linguistic information,the belief interval interpretation of probabilistic linguistic term sets is firstly proposed,which allows the discrete probabilistic linguistic information to be transformed into the continuous form of confidence intervals.This can reduce the uncertainty of information caused by the differentiated knowledge and cognitions of decision makers.Then,two methods of belief interval measures,distance measure and probability measure,are proposed to identify the preference of decision makers.Finally,combining graph theory and Dempster-Shafer evidence theory,the commercial banks' data asset quality evaluation method eliminating the preference of decision makers based on the probabilistic linguistic term sets is constructed.This research applies two theoretical tools that have the advantage of dealing with uncertainty,namely,the neutrosophic soft sets and the probabilistic linguistic term sets,to propose relevant uncertain information measures respectively,and proposes a belief interval interpretation of the probabilistic linguistic term sets to reduce information uncertainty.In addition,the preference of decision makers is divided into preference for criteria and preference for alternatives.Based on the process of preference recognition and elimination,the data asset quality evaluation methods for commercial banks to eliminate preference of decision makers under uncertain environment are constructed.Finally,the effectiveness and superiority of the proposed methods are verified through the analysis and comparison of numerical examples.