Investment Decision-making Models Based On Rough Set Theory

In this thesis, two interactive models were presented to addressinvestment decision-making.First, starting from the standpoint of the role of a manager ininvestment industry, a general decision model was proposed for customerrelationship management (CRM). Then an example was given for theinsurance industry where risks and competitions are high. A new dataprocessing tool, Rough set theory, was used to classify, simplify andextract the data. An advantage of Rough set theory is it does not requirefull understanding and proof of the data, which is becoming veryimportant for modern investment industries since there are manyuncertainties in the customers in this ever-changing society. The datamining technique based on Rough sets could help serve the customers intheir each "life" phases, reveal generalizations, and obtain thesimplest data expression means; it could also help identifycharacteristics of customers, and improve services on a personal basis.Thus, under the aid of data-mining technique, the CRM decision modelcould not only provide effective management for customers, but alsorealize efficient customer-centered services, improve customersatisfaction degree, and develop new customers. Meanwhile, it could beuseful for investment decision-making.On the other hand, a portfolio investment decision model wasestablished for the purpose of choosing potential investment partners.In this model, two mathematical tools, Rough sets and Fuzzy theories arecombined to address incomplete and uncertain data system. Rough settheory is important in retaining data objectivity and Fuzzy set theoryis able to obtain quantitative results from uncertainties. Based on general portfolio investment decision-making rules, Rough set theory wasapplied to make knowledge reduction on the uncertain portfolioinvestment risk decision information system, to analyze the weight ofvarious risk indexes and to explore relevancy in the data studied. Asa result, each weight of evaluation indexes could be obtained. Fuzzycomprehensive evaluation method was finally applied to evaluate theinvestment decision model. The present method proved effective inpractice and provides a useful tool for investors to evade risks indecision-making.The two models studied here differ from each other and yet areinterlocking. With the rapid development of market economy, theprocesses toward multi-structured industrialization and multi-roledinvestment entities are deepened. For instance, an investment industryacts not only as the manager itself, but also an investor of otherindustries. Thus, in an investment industry, both customer relationshipmanagement policy and investment orientation policy are becomingimportant. The present work is in an attempt to provide workable modelsfor investment decision-making and data mining processes.