The paper presents a robust linear programming formulation for separating two disjoint point sets in \( n \)-dimensional real space. The formulation aims to minimize the average sum of misclassified points and guarantees a separating plane when the convex hulls of the sets are disjoint. Unlike previous methods, it does not impose extraneous constraints that can fail in handling certain cases. The authors demonstrate the effectiveness of their method through computational results on the Wisconsin Breast Cancer Database and the Cleveland Heart Disease Database, showing superior performance compared to other linear programming approaches, including Smith's method and a multisurface method. The robustness of the proposed linear program is highlighted, particularly in handling linearly inseparable sets without generating a null solution.The paper presents a robust linear programming formulation for separating two disjoint point sets in \( n \)-dimensional real space. The formulation aims to minimize the average sum of misclassified points and guarantees a separating plane when the convex hulls of the sets are disjoint. Unlike previous methods, it does not impose extraneous constraints that can fail in handling certain cases. The authors demonstrate the effectiveness of their method through computational results on the Wisconsin Breast Cancer Database and the Cleveland Heart Disease Database, showing superior performance compared to other linear programming approaches, including Smith's method and a multisurface method. The robustness of the proposed linear program is highlighted, particularly in handling linearly inseparable sets without generating a null solution.