Furthermore, a changeable update rule was introduced to accelerate the convergence speed, a forget factor and a predefined error bound were used to enhance the robustness of the algorithm.

• 应用遗忘因子和学习律误差容限控制的方法来提高学习算法的鲁棒性。