# Six Sigma: Analysis

Six Sigma Analysis: Asking Why problems occur and where are opportunity to improve them. Finding Y=f(x).
¤ Qualitative Analysis: Process Mapping. Process flow (information, material, money). Cause&Effect- Fishbone. 5 Why’s.
¤ Quantitative Analysis-Statistics: First, data collection then:

•Descriptive: Profiling.
▪Numeric: Mean, deviation, min/max ..
▪Graphic: Histogram. Etc.
•Inferral: point Estimation(mean, etc.). Confidence interval. Estimating distributions: Sample space (possible outcome values: select distinct c1). Random variable: assigning probability value to each possible outcome. Probability distribution: Estimating a curve to fit the probability values. For discrete variables: binomial: high probability, poisson: rare events, etc. For continuous values: Uniform, exponential, Normal, etc.
After that, one can predict the probability as well as percentage of population that of a certain outcome (or set of outcomes) to occur. Also what values x% of the population will fall into. Also we need it in using hypothesis testing(e.g. ANOVA): if the difference between the two dimension values (pricing strategies, before and after a fix, etc) is random (not significant) what is the probability that the values found would occur (randomly)? If too small>>there’s diff! E.g. Fix/solution is good.
Hypothesis testing.

Predictive:
Correlation: helps defining if there is a linear correlation between the problem and a potentially root cause, also a linear correlation between a root cause and a potentially solution measure to optimize.
Regression: helps explaining an advanced (linear or non linear) form or correlation (problem-root cause, root cause-solution measure). It also helps predicting the cost of bad quality: how much loss if bad quality remains, through regression between time and problem CTQ measure.

Levels of Statistical methods:
Collecting Data > Profiling > Descriptive measures > Inference: Experimental design, ANOVA (also:chi-square,t-test): Hypothesis testing > Predictive.