[2018年最新整理]08MultivAnal(多变量分析).ppt

Six Sigma Champion Training Multi-Vari Studies Real World Scenario Manufacturing: A Black Belt or Process Engineer needs to determine the significant variables for a process before conducting a Design of Experiments. Administrative: A Manager needs to determine the family of variation that most effects the picking order of parts to be delivered to the customer. Noise Variables Boxplots  Main Effects Plot for Shift, Piece, and Position Interaction Plot for Shift, Piece, and Position StatQuality ToolsMulti-Vari ChartShiftPiecePosition Data Collection Team must follow sampling plan consistently Note changes that are not part of the normal operation Capture all the data Maintain monitors on gauges Record events that are out of the ordinary Log all data as soon as possible Data Analysis Determine stability and capability Explore relationships among the variables Partition the total variability (variance component analysis) Handling outliers and non-normal data Questions to be answered: What seems to be the major source of variation? Handling Outliers Outliers Boxplots, histograms, dotplots, and moving range charts are good graphic methods for identifying outliers Investigate outliers carefully. They may be data entry errors. They may lead to identifying a critical noise variable. If outliers can be reasonably explained, eliminate them from the data set Procedures for handling outliers Identify graphically If outliers can be explained, then complete analysis as usual. If outliers cannot be explained, analyze the data twice, once with the outliers, once without the outliers. See how the results change and make conditional statements. Handling Non-Normal Data Handling Non-Normal Data Handling Non-Normal Data Handling Non-Normal Data-Weibull Analysis Handling Non-Normal Data-Normal Analysis Example Problem Workshop Example Summary * T R A I N I N G C O N S U L T I N G B u i l d i n g S u c c e s s W o r l d w i d e S O F T W A R E Champion Training Positional