Friedman Benchmark Function dataset 1: Description. This is a synthetic benchmark dataset proposed by Friedman in 1991. The cases are generated using the following method: Generate the values of 5 attributes, X1, ..., X5 independently each of which uniformly distributed over [0.0, 1.0]. Obtain the value of the target variable Y using the equation: y=10(sin(PI)x1x2)+20(x3-0.5)2+10x4+5x5+e where e is a Gaussian random noise N(0,1). 2: Type. Regression 3: Origin. Real world 4: Instances. 1200 5: Features. 5 6: Missing values. No 7: Header. @relation friedman @attribute Input1 real [0.0, 1.0] @attribute Input2 real [0.0, 1.0] @attribute Input3 real [0.0, 1.0] @attribute Input4 real [0.0, 1.0] @attribute Input5 real [0.0, 1.0] @attribute Output real [0.664014955, 28.5903858] @inputs Input1, Input2, Input3, Input4, Input5 @outputs Output