摘要
文章针对汽车电力变压器生产过程中存在的供销不平衡的问题,利用预测模型,对汽车电力变压器的工业生产过程进行优化研究。文章使用Kaggle平台上470条电力变压器故障分析数据,对电力变压器的预期寿命进行回归预测。将GradientBoosting模型与RandomForest等8种模型进行对比,GradientBoosting模型准确率达86%,证明了其预测性能的优越性。此外,文章还对特征进行重要性分析,有助于理解模型的内部工作机制、更好地进行数据预处理和特征工程。
In order to solve the problem of imbalance between supply and marketing in the production process of automobile power transformer,this paper uses a prediction model to optimize the industrial production process of automobile power transformer.In this paper,470 fault analysis data of power transformers on the Kaggle platform are used to regress the life expectancy of power transformers.Comparing the GradientBoosting model with 8 models such as RandomForest,the accuracy of the GradientBoosting model reached 86%,which proved the superiority of its prediction performance.In addition,this paper also analyzes the importance of features,which is helpful to understand the internal working mechanism of the model,and better perform data preprocessing and feature engineering.
作者
张友友
李亮
林梅
林莉
Zhang Youyou;Li Liang;Lin Mei;Lin Li
出处
《时代汽车》
2024年第18期161-163,共3页
Auto Time
基金
2022年四川省大学生创新创业训练项目:基于预测模型的汽车零部件工业流程优化方法研究(107261858)。
关键词
电力变压器
生产流程优化
预测模型
特征重要性排序
机器学习
Power Transformers
Production Process Optimization
Predictive Models
Feature Importance Ranking
Machine Learning
