基于集成学习的煤灰软化温度预测研究

Coal Ash Softening Temperature Prediction Research Based on Ensemble Learning

对煤灰软化温度(ST)的预测任务,现有的传统经验公式及神经网络算法,普遍面临泛用性和泛化能力不足的问题。以堆叠结构组织回归决策森林和AdaBoost回归算法,放松了原始数据规模和特征规模对模型参数规模的限制,提高了模型的泛用性和泛化能力。实验结果表明,该模型预测准确度能够控制在2.25%以内,验证了所提出煤灰软化温度预测方法的有效性。

For the task of coal ash softening temperature(ST)prediction,the existing traditional empirical formulations and neural net work algorithms generally face the problems of insufficient general-ization ability and generalization ability.In this paper,we organize the regression decision forest and AdaBoost regression algorithm with a stacked structure,which relaxes the limitations of the original da-ta size and feature size on the model parameter size,and improves the model's generalization ability and generalization ability.The experimental results show that the prediction accuracy of the model can be controlled within 2.25%,which verifies the effectiveness of the coal ash softening temperature predic-tion method suggested in this paper.