Integrating LSS Principles, Tools, and Techniques in Quality 4.0
Concurrent Sessions (T15 - T20)
Presentation Description: This presentation provides practical guidance how to apply analytics to make decisions based on data that can often be overwhelming, messy, and incomplete. Although the example data come from the manufacturing world, the methods apply equally to the service, healthcare, government, and education realms as well. The presentation addresses the importance of understanding both the problem and the data. It shows ways to address problems such as missing, imprecise, and biased data and gives multiple examples of visualization used to guide analysis. It presents analysis according to both classical statistical methods and more modern machine learning approaches and discusses their relative merits. It emphasizes the importance of clearly communicating the findings.
Six Sigma practitioners have a big toolbox these days for tackling data challenges. This talk employs several of the available analytic tools (e.g., correlation, logistic regression, classification trees, random forests, etc.) using a variety of software tools (e.g., Minitab, Excel, JMP, and Salford Predictive Modeler). Examples come from a real world investigation into a significant manufacturing defect; they tell the story of work done to identify a fix for the defect and demonstrate that the issue was resolved.