Statistical Methods for Process Understanding and Improvement

我的研究为过程改进和理解开发了创新的统计方法。更具体地说，我建议在以下三个领域继续我的工作：统计工程、测量系统评估和过程监控与理解。我的研究结果解决了现实世界的问题，有可能极大地影响工业实践。具体目标包括改进统计工程的定义和扩大应用范围，并基于各种来源的数据，包括制造，最终测试，授权和客户使用，开发适用于Research in Motion的现场故障预测方法。此外，我计划为测量系统的评估和比较制定新的和更好的计划和分析。在这里，一个特别的重点将是放松可疑的假设和处理非标准的测量特征，如计数，有界数据和多变量数据。第三个研究主题侧重于开发模型和方法，以便在输出特性不能直接比较时监测过程，并更好地了解其性能。一个具体的项目涉及开发一种监测手术学习曲线的方法。我也建议重新审视统计过程控制（SPC）的基本概念和术语。当前的环境缺乏连贯性，许多矛盾导致混乱。开发一套一致和简洁的定义将提高对SPC的理解，并有望成功使用。

My research develops innovative statistical methods for process improvement and understanding. More specifically, I propose to continue my work in the following three areas among others: Statistical Engineering, Measurement System Assessment, and Process Monitoring and Understanding. The results of my research which tackle real world problems have the potential to greatly influence industrial practice. Specific goals include improving the definition and widening the application of Statistical Engineering and developing a field failure prediction methodology applicable at Research in Motion, based on data from a variety of sources, including manufacturing, final test, warrant and customer usage. In addition, I plan to develop new and better plans and analyses for the assessment and comparison of measurement systems. Here a specific focus will be on relaxing questionable assumptions and tackling non-standard measurement characteristics, such as, counts, bounded data and multivariate data. A third research theme focuses on developing models and methods for monitoring processes when the output characteristics are not directly comparable and to better understand their performance. One specific project involves developing a method for monitoring the surgical learning curve. I also propose to re-examine the fundamental concepts and terminology of Statistical Process Control (SPC). The current environment has little coherence and many contradictions that result in confusion. Developing a consistent and concise set of definitions will improve SPC understanding and hopefully successful usage.