Quality Prediction Of Complex Mechanical Products Based On Improved DA-XGboost Algorithm

So far,the quality of complex mechanical products has always been concerned by manufacturing companies.Quality prediction in the manufacturing process of complex mechanical products has become a basic method to ensure product quality.However,with the rapid innovation and continuous development of mobile Internet,big data and other emerging industrial information technologies,some new technical problems have emerged in the manufacturing process of complex mechanical products.The research on the quality prediction of complex mechanical products has become the research of domestic and foreign scholars.Hot spots,this article mainly focuses on the research and practice of the quality prediction level of complex mechanical products.The manufacturing mode of complex mechanical products is usually divided into two modes: single and small-scale batch production,and the production process and equipment manufacturing process are complicated,the cost of rework of raw materials and products is too high,and the number of qualified and unqualified products varies greatly.In view of the high dimensionality,small sample and unbalanced characteristics of its quality characteristic data set,this paper proposes a quality prediction model based on the improved DA-XGboost complex mechanical product,which uses the quality data of the complex mechanical product to predict the quality of the product.Solve the quality loss and rework risks caused by the unqualified quality of complex mechanical products.First,according to the characteristics of the quality characteristic data set of complex mechanical products,the Gaussian chaotic mapping algorithm and the binary dragonfly algorithm are combined to propose an improved dragonfly algorithm to identify the quality data set of complex mechanical products,identify the key quality characteristic set,and reduce its quality Feature set,reduce the complexity of quality prediction;secondly,build a quality prediction model for complex mechanical products based on improved DA-XGboost,predict the quality of complex mechanical products,and compare them with simulations and experiments;finally,verify through simulation examples,The improved dragonflyalgorithm used in this paper has been improved in feature selection.The improved algorithm can effectively identify the key quality features of complex mechanical products,improve the stability of features,and reduce the dimensionality of the quality feature data set.In predicting the quality of complex mechanical products,especially for high-dimensional,small sample data sets,the quality prediction model based on CDA-XGboost is better than the quality prediction model based on CDA-SVM.XGboost is more suitable for the quality of complex mechanical products prediction.