Research And Application Of Data Mining Algorithms In The Quality Control Of LCD

After more than ten years of development,Chinese LCD manufacturing has reached a certain stage;LCD related manufacturing technology has matured gradually. However,as Chinese market with the international market's further integration, Chinese manufacturing industry is facing increasingly fierce competition.In order to be invincible,one enterprise has only one choice,improve products' quality.In this case,for LCD manufacturers to make further breakthroughs in product quality there is an urgent need for a suitable method to discover the existence of a variety of small variation in the production process,and an appropriate quality statistical analysis tool to take measures to improve and optimize the process.This study take domestic LCD manufacturer's two quality problem during Array section as analysis goal,collect process and materials data from March 2008 to August 2008,use data mining techniques to explore possible adverse reasons,and establish of forecasting model.The major work and creative points of this paper include:Introduction of the business background and the traditional quality control methods in LCD,take use of Shewhart control chart to control the parameters of every process,prove that every process is in control alone.Base on decision tree and neural network two data mining method to explore the reasons of products' ineligible,proved that even though each of the process is in control alone,there is still space for these two methods to further improve the LCD quality.Compare the result of decision tree model and neural network model.Also,has a certain reference value to other data mining project.