fitting 方法的异常值过滤

training = pd.DataFrame({'x':[3,6,9,15,300, 20,85]}).  原始数据
training_fitting = pd.DataFrame({'x':[4,7,8,30,280, 10,79]})。 原始数据的fitting值，方法不限。xgboost，RF。。。

dif = np.abs(training.x -training_fitting.x) <10
training_data = training[dif].  #过滤高异常的差值。