Applied Research On Control Chart Pattern Recognition Based On Adaboost Algorithm

Quality management is an increasingly prominent role in modern industrial production, the use of statistical process control is an important means of quality management. In many process monitoring analysis and troubleshooting tools, quality control charts is the most classic and effective one. Real-time monitoring of the control chart, Accurate identification of normal and abnormal pattern of control chart,that will be able to find process anomalies, eliminate potential dangers and risks, protect the quality and safety of industrial production, it is of great significance.First, this paper introduced the meaning and the role of quality control charts and control chart pattern recognition, summarized and analyzed the research results made by our predecessors in the field of control chart pattern recognition. On this basis, it proposed a brand new control chart pattern recognition method based on Adaboost algorithm.The article described the algorithm flow and characteristics of Adaboost algorithm, subsequently proposed the optimization strategy of Adaboost algorithm which is applied to the control chart pattern recognition. Then comprehensive and intuitive introduction about the Adaboost algorithm used in this paper intelligent diagnostic procedures design and simulation processes was given, Including a description of the six types of control charts, data simulation process, the generation of training sample set and testing sample set, and highlights based classifier selection (In this paper, Naive Bayesian Classifiers were selected). Finally, through a large number of data simulation instance,it gave the recognition results in the case of different number of iterations and different number of training sample sets, proved that this algorithm is superior to most previous proposed algorithm. On this basis, it gave the effect of recognition in the ordinary case, and improved the algorithm. The results showed that, Adaboot intelligent algorithm application in the field of control chart pattern recognition has unique advantages and excellent performance.