Find best possible first split-point (name of feature and value of split-point) for a regression tree for following data example, where y is target and x1 and x2 are possible features for the split. Use L2 loss as a loss function.

y = (3, 3.5, 4, 6) x1 = (0.1, 0.4, 0.05, 0.2) x2 = (blue, orange, orange, blue)