Analyze Flashcards
(25 cards)
What is the question answered in the Analyze phase?
“What is causing the problem?”
What is the name of the diagram used to collect/find potential root causes?
Ishikawa Diagram or Fishbone diagram
What are the categories of the fishbone diagram?
The 6Ms:
Method, Man, Machine, Material, Measurement, Mother Nature
What is the Tool 3 used for during the Analyze phase?
It is used to build a link between the potential causes (X) with the output measurements (Y), attached to the CTC
During the data analysis, what are the tests that can be used when I have discrete output (Y) and discrete inputs (X)?
Hypothesis tests (proportion):
1-proportion test
2-proportion test
Chi-square
During the data analysis, what are the tests that can be used when I have discrete output (Y) and continuous inputs (X)?
Logistic regression test
During the data analysis, what are the tests that can be used when I have continuous output (Y) and discrete inputs (X)?
Hypothesis Tests:
(Location)
**1-sample t-test
**2-sample t-test
**ANOVA
(Variance/spread)
**2-sample variance
**Test for equal variances
During the data analysis, what are the tests that can be used when I have continuous output (Y) and continuous inputs (X)?
Correlation and regression
During the data analysis, what test that can be used regardless of the type of data the I am working with (discrete or continuous)?
DOE (design of experiments)
What is the p-value?
It is the actual risk taken of rejecting H0, when H0 is actually true.
Why do we do Hypothesis testing?
To verify the potential root causes and make objective decisions about the validity of the assumptions made.
When is the Type I error (alpha error) made?
When we reject the null hypothesis H0, when it is actually correct
When is the Type II error (beta error) made?
When we accept the null hypothesis H0, when it is actually wrong
What is the parameter used to make a decision based on the result of p-value?
if p-value =< 0.05, we reject the null hypothesis H0. This means that there is a difference and it is significant. My data is NOT NORMAL
if p-value > 0.05, we accept the null hypothesis H0. This means that there is no difference, and it is not significant. My data is NORMAL
What is the aim value for Cp and Cpk and Cg and Cgk?
1.3 or higher
what is R-square?
R-square is the coefficient of determination, it explains how much variation is in my model.
R-square is the explanation of known variation of the output through the model
What are the parameters for the Pearson Correlation coefficient?
If r=1 or r=-1; then it is a perfect correlation
If r=0.8 or 0.5<r<0.8; then it is a strong correlation
If r=0; there there is no correlation, “completely independent factors”
What is the VIF and what is the parameter associated with it?
VIF is the Variance Inflation Factor, it checks the independence of the factors in the model.
If the factors depend on each other, they will disturb the model.
if VIF > 5 remove factor from model
if VIF < 5 the factors are independent
What does the confidence interval tell us?
It tell us the range where we expect the median population to be.
What is a residual? from Kahoot
From Kahoot:
It is the difference between the fitted(model) value and the actual value
What are the difference types of designed or planned experiments for Data Analysis?
-Simple Comparison
-OFAT: changing one factor at a time
-Two way ANOVA (for discrete inputs)
-Regression (for continuous inputs)
-Design of Experiments
When do you use DOE?
during Analyze, to find relevant factors of the root causes
during Improve, to find optimal settings for desired goal, to “control” or “steer”
What is the form of the full factorial DOE?
2 to the power of k = basic trials number
-Where 2 is the factor level (min and max, only 2)
-and k is the number of factors
In order to account for the unknown variables, what can we do in the DOE?
-We can repeat the trial, aka short term variation
-we can replicate the trial, aka long term variation
