|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/tabtrainer-minitab-one-sample-poisson-rate/
课程评论：没有评论
**课程名称：** Minitab 中的泊松率假设检验 - Tabtrainer 指南 **课程概述：** 本课程是 Tabtrainer® 认证系列的一部分，旨在帮助学员在实际生产环境中做出统计上可靠的决策。课程将以 Smartboard 公司减震垫生产的真实质量数据为例，教授如何使用 Minitab® 执行单样本泊松率检验。目标是确定观测到的缺陷率是否超过客户设定的严格标准：每批 500 个单位中不允许超过 25 个缺陷。 学员将学习分析 50 个抽样批次的数据，可视化泊松分布，正确表述假设，并解读 P 值、泊松均值和决策边界。课程还将重点讲解在低计数质量环境下，何时应优先选择精确泊松方法而非近似方法。 课程由 TÜV 认证专家、德国 2023 年度最佳教授 Murat Mola 博士主讲，通过精益六西格玛思维、Minitab 实际工具和工厂一线的案例分析，展示如何将统计理论转化为对业务至关重要的建议。 **学习内容：** 学员将深入了解一个现实的生产质量控制场景，即 Smartboard 公司减震垫的生产。具体目标是评估生产过程是否满足客户的严格要求：每批 500 个减震垫最多允许 25 个表面缺陷，相当于 5% 的缺陷率。 由于对每个减震垫进行检查在经济上不可行，学员将使用包含 50 个随机抽样批次的数据集。每个批次包含 500 个零件，缺陷数量通过自动表面检测系统进行测量。在此基础上，学员将运用单样本泊松假设检验进行统计分析。 完成本课程后，学员将能够： * 理解生产环境中质量控制的背景和重要性，尤其是在客户有严格规格的情况下。 * 处理真实的样本数据，并解读其结构，包括批次号、样本量和检测到的缺陷数。 * 学习关键统计术语，如： * **总发生次数 (Total Occurrences)：**所有样本的总缺陷数。 * **样本率 (Sample Rate)：**每个零件的平均缺陷数。 * **样本均值 (Sample Mean)：**每个批次的平均缺陷数。 * 理解为什么泊松分布是模拟此类缺陷数据的合适选择，并了解其与二项分布、正态分布和卡方分布的区别。 * 可视化和解释泊松概率分布，并基于均值 (λ 或 μ) 理解其参数化。 * 执行假设检验，以估计总体缺陷率并评估过程是否仍在控制之中。 * 正确表述假设： * **原假设 (H₀)：**每个批次的平均缺陷数 ≤ 25。 * **备择假设 (H₁)：**每个批次的平均缺陷数 > 25。 * 选择精确泊松方法而非正态近似，因其在低计数情况下具有更高的准确性和更好的区分度。 * 解读假设检验结果，包括： * 计算出的泊松率 (λ) 和均值 (μ)。 * P 值及其对决策的影响。 * 认识样本值与总体估计值之间的差异，以及假设检验如何弥合这一差距。 * 做出数据驱动的质量管理决策：即使样本看起来略高于客户阈值，检验也可能显示过程在 95% 置信度下仍然统计可控。 总而言之，学员将能够向管理层提供知情且统计上可靠的建议，决定是需要立即改进流程，还是在当前流程表现接近关键限制时仍可接受。 本单元展示了精益六西格玛工具，如泊松分布和假设检验，如何在实际生产环境中应用，将统计理论与业务影响相结合。
Welcome to the Tabtrainer® Certified Series - your go-to platform for statistically sound decision-making in real production environments.In this training unit, you'll learn to apply the One Sample Poisson Rate Test using Minitab®, based on real quality data from shock pad production at Smartboard Company. The goal: determine whether the observed defect rate exceeds the customer's strict threshold of 25 defects per batch of 500 units.You'll analyze 50 sampled batches, visualize Poisson distributions, formulate hypotheses correctly, and interpret p-values, Poisson means, and decision boundaries. You'll also understand when to use the exact Poisson method instead of approximations - a key point in low-count quality environments.Taught by Prof. Dr. Murat Mola, TÜV-certified expert and Professor of the Year 2023 in Germany, this course shows how to translate statistical theory into business-critical recommendations - using Six Sigma thinking, real Minitab tools, and practical casework from the factory floor.Learning DescriptionIn this training unit, students are introduced to a realistic quality control scenario in the shock pad production at Smartboard Company. The goal is to assess whether the production process meets the customer's strict requirement: a maximum of 25 surface defects per batch of 500 shock pads, equivalent to a 5% defect rate.Since inspecting every shock pad would be economically unfeasible, students work with sample data consisting of 50 randomly selected batches. Each batch contains 500 parts, and the number of defects per batch was measured using an automatic surface inspection system. Based on this data, the students perform a statistical analysis using the one-sample Poisson hypothesis test.By completing this unit, students will learn to:Understand the background and relevance of quality control in a production environment with tight customer specifications.Work with real sample data and interpret its structure, including batch numbers, sample sizes, and detected defects.Learn key statistical terms:Total Occurrences - the total number of defects in all samples combinedSample Rate - the average number of defects per single partSample Mean - the average number of defects per batchUnderstand why the Poisson distribution is the appropriate choice for modeling such defect data, and how it compares to the Binomial, Normal, and Chi-square distributions.Visualize and interpret the Poisson probability distribution and understand its parameterization based on the mean (λ or μ).Perform a hypothesis test to estimate the population defect rate and assess whether the process is still in control.Learn the correct formulation of:Null Hypothesis (H₀): The average number of defects per batch is ≤ 25.Alternative Hypothesis (H₁): The average number of defects per batch is > 25.Select the exact Poisson method over the normal approximation due to its higher accuracy and better selectivity in low-count situations.Interpret the results of the hypothesis test, including:The calculated Poisson rate (λ) and mean (μ)The p-value and its implications for decision-makingRecognize the difference between sample-based values and population-based estimations, and how a hypothesis test can bridge this gap.Make data-driven quality management decisions:Even if the sample appears just above the customer threshold, the test might show the process is statistically still in control-with 95% confidence.In conclusion, students will be able to make informed, statistically sound recommendations to management-deciding whether immediate process improvements are necessary or if the current process performance is acceptable despite being close to the critical limit.This unit demonstrates how Six Sigma tools like the Poisson distribution and hypothesis testing are applied in real-world production environments, combining statistical theory with business impact.
