6SIGMA工具Categorical Output Analysis.pptVIP

Breakthrough Management Group Version 1.0 Process Evaluation with Categorical Data Module Objectives Review Tools to evaluate Y=f(X) when Y is: X is: One factor, one level One proportion test Sign test Rank sum test One factor, two levels Two proportion test One factor, multiple levels Analysis of Means Chi-Square Goodness of Fit Logistic Regression The Master Black Belt as Expert Throughout your organization, project leaders will be asking for your help to plan and execute data analysis Many of these requests will involve the manipulation of categorical data, for both input and outputs Tools for continuous data: Means testing Variance testing Shape testing Are not applicable !!!!! Other tools specially geared for categorical data are needed Categorical vs. Variable Data Nature of Transactional Projects trans·ac·tionFunction: noun1 a : something transacted; especially : an exchange or transfer of goods, services, or funds 2 a : an act, process, or instance of transacting b : a communicative action or activity involving two parties or things that reciprocally affect or influence each other- trans·ac·tion·al /-shnl, -sh-nl/ adjective Many Supporting Operational Processes Measuring Transactional Success The quality of the TRANSACTION is usually measured two ways: Cycle Time Quality Score Good/Bad Good/Marginal/Bad Great/Good/Neutral/Bad/Horrible Determining Significant Inputs Following the Measure Phase, data for inputs and outputs describing distinct “cases” is available for analysis. For Example: Analysis is required to determine predictive relationships between inputs and outputs Definitions Attribute Data: Counting the number of times a certain event occurs Count usually for a specified period of time Rates Count per time, number of units, surface area Proportions Divides a group into different piles, reports the count in a pile per total in group Y=f(X) X is Single Factor, One Levels Consider cases where: Y is a binary measure for