安氏亚太第五讲QFD`.ppt

三、非参数检验 1、符号检验 小样本情形 设立假设：H0：P=0.05 H1:P≠0.5 设a=0.05 0.192 经简化后实际应用 H0：P=0.05 H1:P≠0.5 其中P是出现“+”号概率 调查结果为，喜欢A产品为2，喜欢B产品为10， 设a=0.05，另也可设 0.01,0.05,0.10,0.25 结果如何判断？ 若只是单边（考虑低端或高端），则将上述水平减半，查表时不变。 大样本情形 HO：P=0.05 且样本大于30，概率的分布近似正态 公司推出A、B两款新手机，对200名拥有该手机的经销商进行满意度调查，72位经销商认为A手机较好，103位经销商认为B手机较好，25人未给出意见，这项市场调查是否显示出A、B两款手机存在着显著差异。 H0：P=0.5 H1：P≠0.5 ，其中P是来自对A手机的评价 利用上述公式，我们可以进行解答 2、中位数的符号检验 中位数将一个总体分割成两部分，且总体中50%的值将大于或等于中位数，而50%的值将小于或等于中位数。 我们可以采用符合检验，把样本中数据大于假设的中位数的值用加号，小于假设的中位数的值用减号，任一恰好等于假设的中位数的数据应被剔除。 adc * adc * adc * adc * Several systems exist for quantifying potential failure modes in order to drive corrective actions. The actual system used depends on the needs of the organization. The following is typical of many of the FMEA analysis systems in use today and should be customized to meet the specific needs of Seagate. Three items are scored as follows: 1. Severity: assessment of the failure effects on the local area, next level, and end user. Severity rating applies to the effects and the effects only. A change in severity rating can only be achieved through design or application change. 2. Occurrence: The likelihood that a specific cause will occur and will result in the specific failure mode considering: * history / judgment - is design, application or end use a carry-over from a previous product? * significance of changes from the previous design and application * environmental conditions * customer expectations 3. Effectiveness / Detection: An assessment of the ability of the current / proposed controls to identify the failure mode prior to occurring or prior to release to production, or prior to shipment to the customer. This rating can be improved only by improving the control system or the design. Each of the above is typically scored on a scale of 1 - 10 with a 1 being desirable (low occurrence, low severity, high detection). The forth measure is the product of the three and is referred to as the Risk Priority