Estimation is associated with Fear, Uncertainty and Death marches. Most of us would rather not estimate. Yet, sometimes we do need estimates and commitments, even on "estimation-less" projects. Play a series of estimation games to experience how different techniques deliver very different results. Learn a few simple rules that turn you into a reliable estimator. But correct estimates aren't enough. See what else is required to deliver on your promises. Learn to deal with the destructive games people play with estimates. Estimating can be Fun, embracing Uncertainty and Delivering.
Introduction, Terminology and concepts, Introduction to statistics, Central tendencies and distributions, Variance, Distribution properties and arithmetic, Samples/CLT, Basic machine learning algorithms, Linear regression, SVM, Naive Bayes
Estimation is associated with Fear, Uncertainty and Death marches. Most of us would rather not estimate. Yet, sometimes we do need estimates and commitments, even on "estimation-less" projects. Play a series of estimation games to experience how different techniques deliver very different results. Learn a few simple rules that turn you into a reliable estimator. But correct estimates aren't enough. See what else is required to deliver on your promises. Learn to deal with the destructive games people play with estimates. Estimating can be Fun, embracing Uncertainty and Delivering.
Introduction, Terminology and concepts, Introduction to statistics, Central tendencies and distributions, Variance, Distribution properties and arithmetic, Samples/CLT, Basic machine learning algorithms, Linear regression, SVM, Naive Bayes
2. Matrimonial Unrest
• Mr. & Mrs. Hartford
– Newly married
– Having numerous fights
– Unsure as to the cause(s)
– Visits a team of a psychologist and a marriage
counselor
– Collect data of various events that may trigger
fights
3. Matrimonial Unrest
• Potential factors for a fight:
– Number of calories consumed
– Number of hours of sleep the night before
– Bad day at work
– School the next day
– Number of cups of coffee consumed that day
– Uncomfortable symptoms of PMS that day
– Bills paid that day
4. Matrimonial Unrest
• Hypothesis Testing
– Define the hypothesis
• Null hypothesis
– States the obvious
• Alternative hypothesis
– Bears the burden of proof
– Calculate the test statistic
– Make a decision
• If the sample mean is close to the assumed population
mean, do not reject the null hypothesis
• If the sample mean is far from the assumed population
mean, reject the null hypothesis
5. One Sample Hypothesis Testing
n > 30
• : =AVERAGE()
• : =STDEV()
• zOBS : =STANDARDIZE(x, , /SQRT(n))
• zCRIT : =NORMSINV(1- /# of tails)
• p-value : =(# of tails) * (1-
NORMSDIST(ABS(zOBS)))
6. One Sample Hypothesis Testing
n < 30
• : =AVERAGE()
• s : =STDEV()
• tOBS : =STANDARDIZE(x, , s/SQRT(n))
• tCRIT : =TINV(2 /# of tails,df)
• p-value : TDIST(ABS(tOBS)),df, # of tails)
7. Not Enough Sleep
• Null: – Not enough sleep (less than
7 hours) does not lead to more fights.
• Alternative: – Not enough sleep (less
than 7 hours) does lead to more fights.
8. Not Enough Sleep
• Hypothesized Value = .267 • zOBS = (.073)
• Sample Mean = .262 • zCRIT = 1.645
• Standard Deviation = .445 • P-value = .471
• Sample Size = 42 • -level = .05
• Do not Reject the Null
9. Not Enough Sleep
• Null: – Not enough sleep (less than
6 hours) does not lead to more fights.
• Alternative: – Not enough sleep (less
than 6 hours) does lead to more fights.
10. Not Enough Sleep
• Hypothesized Value = .267 • tOBS = (.152)
• Sample Mean = .250 • tCRIT = 1.753
• Standard Deviation = .447 • P-value = .441
• Sample Size = 16 • -level = .05
• Do not Reject the Null
11. Bills Paid That Day
• Null: – Bills paid that day does not
lead to more fights.
• Alternative: – Bills paid that day does
lead to more fights.
12. Bills Paid That Day
• Hypothesized Value = .267 • tOBS = .313
• Sample Mean = .333 • tCRIT = 2.015
• Standard Deviation = .516 • P-value = .383
• Sample Size = 6 • -level = .05
• Do not Reject the Null
13. PMS Symptoms That Day
• Null: – PMS Symptoms that day
does not lead to more fights.
• Alternative: – PMS Symptoms that
day does lead to more fights.
14. PMS Symptoms That Day
• Hypothesized Value = .267 • tOBS = 2.041
• Sample Mean = .600 • tCRIT = 1.833
• Standard Deviation = .516 • P-value = .036
• Sample Size = 10 • -level = .05
• Reject the Null
15. A Bad Day at Work
• Null: – A bad day at work does not
lead to more fights.
• Alternative: –A bad day at work does
lead to more fights.
16. A Bad Day at Work
• Hypothesized Value = .267 • tOBS = 1.680
• Sample Mean = .500 • tCRIT = 1.771
• Standard Deviation = .519 • P-value = .059
• Sample Size = 14 • -level = .05
• Do not Reject the Null
17. Too Much Coffee
• Null: – Too much coffee (2+) does
not lead to more fights.
• Alternative: – Too much coffee (2+)
does lead to more fights.
18. Too Much Coffee
• Hypothesized Value = .267 • tOBS = .925
• Sample Mean = .364 • tCRIT = 1.721
• Standard Deviation = .492 • P-value = .183
• Sample Size = 22 • -level = .05
• Do not Reject the Null
19. School the Next Day
• Null: – School the next day does not
lead to more fights.
• Alternative: – School the next day
does lead to more fights.
20. School the Next Day
• Hypothesized Value = .267 • tOBS = .314
• Sample Mean = .300 • tCRIT = 1.729
• Standard Deviation = .470 • P-value = .378
• Sample Size = 20 • -level = .05
• Do not Reject the Null
21. Conclusions
• As the psychologist
– Not enough sleep (7 hours)
• |OBS|< CRIT and P-value > α-level
We do not reject the null hypothesis, meaning less than 7
hours of sleep does not trigger the fights.
– Not enough sleep (6 hours)
• |OBS|< CRIT and P-value > α-level
We do not reject the null hypothesis, meaning less than 6
hours of sleep does not trigger the fights.
22. Conclusions
• A bad day at work
• |OBS|< CRIT and P-value > α-level
We do not reject the null hypothesis, meaning that a bad day
at work does not necessary lead to a fight.
• Bills paid that day
• |OBS|< CRIT and P-value > α-level
We do not reject the null hypothesis, meaning that when bills
were paid that day it does not necessary lead to a fight.
23. Conclusions
• School the next day
• |OBS|< CRIT and P-value > α-level
We do not reject the null hypothesis, meaning that having
school the next day does not lead to a fight.
• As a marriage counselor
• Too much coffee (2 +)
• |OBS|< CRIT and P-value > α-level
We do not reject the null hypothesis, meaning that drinking
too much coffee does not lead to a fight.
24. Conclusions
• PMS symptoms
• |OBS|> CRIT and P-value < α-level
We reject the null hypothesis, meaning that there is enough
evidence to say that PMS symptoms lead to fights.
25. Recommendations
• Talk to a specialist on how to deal with PMS
symptoms.
• More thorough analysis of the relationship.
• Combination of 2 or more factors may be the
cause so more calculations are needed.
• As a last resort: they are not meant for each
other >>>> Get a good lawyer!!!