Abstract: Usually it is taken for granted to achieve the maximal profit at the cost of the minimal risk. It is an important problem of how to balance profit and risk, considering introducing profit and risk to attribute reduction so as to find practical algorithms in decision-making process. A decision-theoretic model was built, which could balance profit and risk combining with decision-theoretic rough set model and minimum risk of Bayes decision and find optimal combinations of risk in a certain level of expected profit, then a heuristic algorithm of attribute reduction was proposed, which took the function of balancing profit and risk as the target of heuristic attribute reduction, and it could reduce the scales and complexity of data model, and then improve the simulation precision of the model system.

Key words: attribute reduction, decision table, rough set, profit, risk