| 边坡稳定性自动机器学习预测方法研究* |
| |
| 引用本文： | 张化进,吴顺川,张中信,孙俊龙,韩龙强.边坡稳定性自动机器学习预测方法研究*[J].中国安全生产科学技术,2023,19(1):35-40. |
| |
| 作者姓名： | 张化进 吴顺川 张中信 孙俊龙 韩龙强 |
| |
| 作者单位： | （1.昆明理工大学 国土资源工程学院，云南 昆明 650093；2.自然资源部 高原山地地质灾害预报预警与生态保护修复重点实验室，云南 昆明 650093） |
| |
| 基金项目： | * 基金项目: 国家重点研发计划项目(2017YFC0805303)；云南省创新团队项目(202105AE160023) |
| |
| 摘 要： | 为了简便有效地评估边坡稳定性状态，针对目前传统机器学习的算法选择与超参数优化等难题，提出1种边坡稳定性自动机器学习预测方法。首先，简要介绍5种主流开源自动机器学习框架；其次，以422组边坡稳定性样本为数据集，进行自动机器学习纯自动化训练，并与传统机器学习对比分析模型的性能与耗时；最后，综合讨论与比较典型自动机器学习框架的特性。研究结果表明:自动机器学习预测效果总体上优于传统机器学习，提升边坡稳定性预测准确率和稳健性，且无需人为干预。研究结果可为岩土工作人员准确可靠地评价边坡稳定性提供便捷条件。
|
| 关 键 词： | 边坡工程 稳定性预测 机器学习 自动机器学习 |
Research on automatic machine learning prediction method of slope stability |
| |
| ZHANG Huajin,WU Shunchuan,ZHANG Zhongxin,SUN Junlong,HAN Longqiang.Research on automatic machine learning prediction method of slope stability[J].Journal of Safety Science and Technology,2023,19(1):35-40. |
| |
| Authors: | ZHANG Huajin WU Shunchuan ZHANG Zhongxin SUN Junlong HAN Longqiang |
| |
| Affiliation: | (1.Faculty of Land Resource Engineering,Kunming University of Science and Technology,Kunming Yunnan 650093,China;2.Key Laboratory of Geohazard Forecast and Geoecological Restoration in Plateau Mountainous Area,Ministry of Natural Resources of the People’s Republic of China,Kunming Yunnan 650093,China) |
| |
| Abstract: | In order to evaluate the slope stability state easily and effectively,an automatic machine learning prediction method of slope stability was proposed aiming at the problems of algorithm selection and hyperparameter optimization faced by the current traditional machine learning.Firstly,5 mainstream open source automatic machine learning frameworks were briefly introduced.Secondly,422 groups of slope stability samples were taken as the data set to carry out the pure automatic training of automatic machine learning,and the performance and time consumption of the model were compared with traditional machine learning.Finally,the characteristics of typical automatic machine learning frameworks were comprehensively discussed and compared.The results showed that the prediction effect of automatic machine learning was generally better than traditional machine learning,improved the prediction accuracy and robustness of slope stability,and did not require the human intervention.The research results can provide convenient conditions for geotechnical staff to evaluate the slope stability accurately and reliably. |
| |
| Keywords: | slope engineering stability prediction machine learning automatic machine learning |
|
| 点击此处可从《中国安全生产科学技术》浏览原始摘要信息 |
|
点击此处可从《中国安全生产科学技术》下载全文 |
