无监督机器学习驱动的飞机备件分类方法
Unsupervised Machine Learning-Driven Classification Method for Aircraft Spare Parts
为挖掘飞机备件保障规律,提出一种无监督机器学习驱动的飞机备件分类方法。通过最大信息系数检测飞机备件保障数据各维度相关性,采用基于高斯核函数的核主成分分析(KPCA)预处理相关性低的飞机备件保障数据,应用牛顿-拉夫逊优化算法(NRBO)和动态模糊参数寻找飞机备件模糊C均值(FCM)聚类质心最佳位置,自适应迭代生成飞机备件分类最优结果。实验结果表明,在相同飞机备件保障主成分数据条件下,相较于传统模糊C均值聚类、遗传算法优化模糊C均值聚类、粒子群优化模糊C均值聚类3种方法,该方法拥有更优越的快速探寻收敛性能和跳出局部最优解能力,可实现更佳效果的飞机备件分类,为飞机备件采购、库存、修理等保障决策提供科学依据。
To explore the support patterns of aircraft spare parts, an unsupervised machine learning-driven classification method for aircraft spare parts is proposed. Firstly, the maximal information coefficient is used to detect the correlation between various dimensions of aircraft spare parts support data. Secondly, kernel principal component analysis (KPCA) based on the Gaussian kernel function is adopted to preprocess the aircraft spare parts support data with low correlation. Finally, the Newton-Raphson-based optimizer (NRBO) and dynamic fuzzy parameters are applied to find the optimal position of the fuzzy C-means (FCM) clustering centroid for aircraft spare parts, and the optimal result of aircraft spare parts classification is generated in an adaptive iterative manner. Experimental results show that, under the condition of the same principal component data for aircraft spare parts support, compared with 3 traditional methods, FCM clustering, genetic algorithm-optimized FCM clustering, and particle swarm optimization-based FCM clustering, this method exhibits superior performance in rapid convergence exploration and the ability to escape local optimal solutions. It can achieve better aircraft spare parts classification results and provide a scientific basis for support decisions related to aircraft spare parts, such as procurement, inventory management, and maintenance.
无监督机器学习 / 飞机备件分类 / 核主成分分析 / 牛顿-拉夫逊优化算法 / 模糊C均值聚类
unsupervised machine learning / aircraft spare parts classification / kernel principal component analysis / Newton-Raphson-based optimizer algorithm / fuzzy C-means clustering
| [1] |
王桢,黄俊杰,易朋兴.基于改进FCM算法的舰船备件分类方法研究[J].火力与指挥控制,2024,49(10):25-33. |
| [2] |
王强,贾希胜,程中华, |
| [3] |
赵青雨,苏之昀,夏唐斌, |
| [4] |
杨华强,尹亮,赵青雨, |
| [5] |
刘晔,刘晓.基于数据分析的汽车备件分类策略研究[J].工业工程与管理,2018,23(3):80-86. |
| [6] |
舒健生,赖晓昌,李亚雄, |
| [7] |
雷光远,张涛,唐永聪, |
| [8] |
崔秀芳,林浩涛,安楠楠, |
| [9] |
郭俊霞,谢自力,毛申申, |
| [10] |
|
| [11] |
鲁明智,张家明,邱培城, |
| [12] |
杜文斌,王伟.基于NRBO优化的BP神经网络草莓农残检测系统[J].国外电子测量技术,2024,43(8):189-197. |
| [13] |
王敏,何腾松,彭鼎.不同FCM聚类算法的模糊时间序列预测模型比较[J].数学的实践与认识,2024,54(5):128-140. |
| [14] |
郭喜峰,孟铭,宁一, |
| [15] |
许彩芳.基于PSO-FCM智能算法的计算机网络入侵检测方法[J].佳木斯大学学报(自然科学版),2022,40(1):60-62;68. |
| [16] |
余浩,高镱滈,潘险险, |
空军装备综合研究自主安排一般项目(KJ2023C010120-7)
/
| 〈 |
|
〉 |