Abstract:

 Fuction similarity degree based on intervals and discrete points is proposed, and the method of fuction clustering based on the two varied function similarity degree is given. Using the classifications of function universe,fuction rough sets is generalized, and ε-function rough sets is put forward. The investment model in economic system is built, and the solution to choose investment portfolio is presented using the tool of ε-function rough sets.

Key words: ε-function rough sets; function similarity degree; function clustering; classification precision; identification precision