Abstract:  A bond recommendation system based on users risk preference is proposed and designed, and content-based recommendation algorithm is used to obtain users risk preference score through AHP (analytic hierarchy process) and fuzzy comprehensive evaluation method, to evaluate the default risk of bonds by using modified KMV model, to construct the recommendation strategy model of “the same kind is better” according to the risk matching principle, and finally to realize the simulation case analysis and verify the availability and effectiveness of the bond recommendation system.

Key words:  bond recommendation, fuzzy comprehensive evaluation method, risk preference, KMV model