Mark J. Anderson 報告者：張育群

Presentation on theme: "Mark J. Anderson 報告者：張育群"— Presentation transcript:

1 Mark J. Anderson http://www.statease.com/pubs/breaddoe.pdf 報告者：張育群
Screening Ingredients Most Efficiently with Two-Level Design of Experiments Mark J. Anderson 報告者：張育群

2 Screening ingredients for making bread
A： Liquid Water(-) or Milk(+) B： Oil Butter(-) or Margarine(+) C： Flour Regular(-) or Bread(+) D： Yeast Regular(-) or Bread(+)

3 Outline-Experiment 製造麵包-參考食譜將A、B、C、D的各個狀 態再加上適當的salt & sugar隨機的放
入烤麵包機中 測試團評分-由3個女兒、 1個交換學生、太太 組成(由於口感不佳，因此將食譜 中加入2茶匙的糖改為2湯匙)

4

5 Unexpected conclusion
在某些情況下會製造出失敗的麵包 麵包未蓬鬆 麵包中仍殘有麵糊 重複狀態1、2、3再製造一次麵包 (1-成功、0-失敗)

6

7

8 Remark In this case the only alias of concern was AC=BD, but since neither the parents (B and D) of the BD interaction came out statistically significant,and these factor (oil and yeast) seemed unlikely to interact.

9

10 Semifold (IV) Lay out the opposite single factor foldover from the original data Perform only half of the foldover rums by selecting those where the chosen factor is either at its low level or high level.

11 Std A B C D E F AC BE 1 － ＋ 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

12 Std A B C D E F AC BE 17 － ＋ 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

13

14

15 Unfortunately Due to the unexpected failures in getting my bread to rise, I lost sight of my original objective. Turned for further revelations on mysteries of making tasty product from bread-making machines

16 END 謝 謝 大 家

17 Reference Anderson, M.; Whitcomb, P.; How To Save Runs, Yet Reveal Breakthrough Interactions, By Doing Only A Semifoldover On Medium-Resolution Screening Designs, th Annual Quality Congress of the American Society of Quality (Milwaukee, WI.).

18