Midterm - Summary of everything from the first half of ISOM 350 course with Steve Walton PDF

Preview

Summary

Summary of everything from the first half of ISOM 350 course with Steve Walton ...

Description

ISOM 350

The magical isom study guide of wisdom  Types of probability: Simple/marginal - the probability a person has blonde hair P(B) Joint - the probability that a person has blonde hair and green eyes P(B and G) Conditional - the probability a person has green eyes given that they have blonde hair P(G/B)

VUCA: volatility, uncertainty, complexity, ambiguity (matches decision type to context) Decision types Obvious: “known knowns” decisions where results are known Volatility: “known knowns, within a range” they are single dimensional, you know how to calculate/get data, but only know the range and not exact #’s Uncertainty: “known unknowns” they’re multi-dimensional and you know the “rules of the game” Ambiguity: “unknown unknowns” uncertain about uncertainty, don’t know what questions to ask Levels of Data: There  are four types of data in descriptive statistics ● Qualitative data - reflect non-numeric issues 1. Nominal data puts data into categories, but no sense of order (zip code, ethnicity, political party) 2. Ordinal data adds order, but no sense of magnitude (freshman, sophomore, junior, senior) ● Quantitative data - reflect numeric issues 1. Interval data adds magnitude, but typically are man-made scales (Dow Jones Industrial Average, temperature) 2. Ratio data adds a "real zero"

ISOM 350

4 categories of descriptive statistics: Measures of central tendency - answers the question "what is typical?"  (MEAN, MEDIAN, MODE) Measures of dispersion - answers the question "how far is a data point from typical?" (RANGE, IQR, STANDARD DEVIATION) The interquartile range - is calculated as Q3 - Q1. Measure of comparison - answers the question "how unexpected is a data point?"  (Z SCORE) Measure of association- answers the question “how do two variables move together?” (correlation coefficient, coefficient of variance)

5 number summary: min, max, median, Q1, Q3 Technical: "the mean number of beds in the data set is 209.9, with a median of 160. Because the mean is greater than the median, the data are skewed. This suggests that there are extreme values in the data." Managerial: “a typical hospital in the geographic area we studied has 210 beds. However, half of the hospitals in this area have fewer than 160 beds. This suggests that there are some hospitals that are significantly larger than the average hospital." Three ways we generate probability: 1. An empirical probability - one that comes from data (a person booked on a flight does not show up) 2. A subjective probability - one that comes from a person's informed opinion (adding a second flight will be profitable)

ISOM 350

3. A p riori, theoretical or classical probability - one that comes from prior knowledge (flipping a coin) Mutually exclusive: P(A and B) = 0, means that two events cannot occur simultaneously; the probability of it is 0 Independence: P(A/B) = P(A), the outcome of an event, event B, doesn’t impact the probability of A occuring. Collectively exhaustive: P(A) + P(~A) = 1 Mutually exclusive: P(A and B) = 0 Total Probability. T  he first rule is total probability. We described this in class as "everything has to add up." All of the rows of a contingency table have to add up, so do all the columns.

Define cross-tabulations of categorical data: Cross-tabs can present frequency as a count, frequency as a percentage, or the cumulative frequency. Typically, cumulative frequency is best used when you have ordinal data. Explain why standard deviation requires squaring the difference between the observed value and the mean. ? Explain why standard deviation requires dividing by n-1 instead of just by n. ? Standard Deviation ● Is the typical distance individuals are from the mean ● Divide by n-1 instead of n so it is an unbiased estimator, so it is good estimator of the population standard deviation. Z-Score ● It’s the number of standard deviations an individual is away from the mean ● It’s a measure of comparison ● (xi- xbar)/SD

Explain the ideas of central tendency and dispersion; include similarities and differences.

Add A New Piece of Information -

Hardly ever effect median. It may change it by a couple of points but not by anything significant If a number is over the mean, the mean will increase, if its under the mean the mean will decrease

ISOM 350

-

If a new piece of information is less than 1 standard deviation above/below the mean, it will decrease the standard deviation. If it’s more than 1 standard deviation away from the mean, it increases the standard deviation

Technical analysis (central tendency) ● Mean (the typical) median (50% is less/more than this value) ○ define by themselves and then compare these two, if mean if larger than median there is skewed data to the right because of extreme values ● mode, min, max ● Standard deviation (on avg hospitals are 50,000 away from the mean) range (big or small) ● Empirical rule ○ 68% of data falls between 1 sds ○ 95% 2 sds ○ 99.7 is 3 sds ● Outliers ○ Z-Score of min/max (if absolute value >3, its an outlier) ○ Is this an extreme value?

Business Analysis (managerial) -

Mean vs Median (artificially inflated) Mode (most frequent) Number of people in survey (Out of 261 hospitals,.....) SD Max/Min (are these outliers?)

Out of 261 firms, the typical sales were $25,176.10 (millions). 50% of the firms experienced sales less than $1863.95 million. This shows that there are a few firms artificially inflating the typical sales that we see. This is further explained by the outlier of the maximum value of 263901.53, which is extremely unusual. The minimum the value really isn’t that far away the mean, so it is not really of concern.

To do: Go through exams

ISOM 350

Do a full technical analysis Learning objectives...