引導式測驗 · 中等 等級
統計過程管制圖測驗
測試 Shewhart、CUSUM 和 EWMA 訊號、控制與規格限、穩定基線、小偏移、調查和 ML 監控差距。
問題 1 的 8
What does a point beyond a control-chart limit establish by itself?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- What does a point beyond a control-chart limit establish by itself?
- The exact machine component that failed
- Proof that the product violated every customer specification
- A signal to investigate a departure from the modeled baseline, not its cause
- A reason to recalculate limits immediately around the new point
- How do control limits differ from product specification limits?
- They are always numerically equal when a process is stable
- Control limits describe expected process variation under a baseline; specifications define acceptability
- Control limits are legal quality standards set by the buyer
- Specification limits are calculated only by an EWMA recursion
- Why can a small persistent mean shift escape a basic Shewhart chart for a while?
- Shewhart charts remove all historic observations
- A stable baseline prevents any change in the plotted process
- The chart always uses a cumulative sum instead of individual points
- Individual plotted statistics can remain inside limits despite repeated small departures
- What feature makes CUSUM useful for detecting sustained small changes?
- It ignores every observation except the latest one
- It accumulates departures from a target across successive observations
- It automatically identifies the cause of drift
- It guarantees no false alarms for a stable process
- In an EWMA chart, how are older measurements represented?
- They receive diminishing weights through the recursive average
- All older values receive the same weight as the newest value
- They are deleted before each calculation
- Only the oldest measurement is used as the baseline
- What should a team establish before interpreting an SPC signal?
- A stable reference period, statistic, sampling plan and alert rule
- Only the most recent outlying value
- A new baseline recalculated after each alarm
- A chart color scheme that labels all points acceptable
- A model's input-feature chart is stable, but outcome labels are not yet available. What remains unverified?
- Whether the input pipeline returned any numbers
- Whether a control chart has a center line
- Whether the feature distribution stayed near the monitored baseline
- Whether task performance on real outcomes changed
- After a documented, stable process redesign, how should old control limits be handled?
- Move them automatically after every alarm
- Treat them as permanent even when the process definition changed
- Deliberately establish and document an appropriate new baseline
- Discard all earlier monitoring records without an audit trail