Uploaded by lekieah.williams

Midterm Review for Business Quant

advertisement
Midterm Review for Desc
is
ar stu
ed d
vi y re
aC s
o
ou urc
rs e
eH w
er as
o.
co
m
Chapter1
Which type of management system includes the integration of technology and
software across an organization?
 Enterprise resource planning systems
Which of the following best characterizes an independent variable?

Function value depends upon its value
Th
Which one of the following indicates the two broad decision categories?
 Programmed and non-programmed
Which of the following categories of modeling techniques provide a course of action
to the decision maker?
 Prescriptive models
Which one of the following describes how a decision-maker views or perceives a
problem?
 Framing
sh
Anchoring effects arise when a seemingly trivial factor serves as a starting point (or
anchor) for estimations in a decision-making problem. Decision makers adjust their
estimates from this anchor but nevertheless remain too close to the anchor and usually
under-adjust.
Framing effects refer to how a decision maker views or perceives the alternatives
in a decision problem—often involving a win/loss perspective.
This study source was downloaded by 100000831008184 from CourseHero.com on 10-12-2021 14:55:22 GMT -05:00
https://www.coursehero.com/file/38690366/Midterm-Review-for-Descdocx/
Chapter 11
is
ar stu
ed d
vi y re
aC s
o
ou urc
rs e
eH w
er as
o.
co
m
Horizontal pattern—stationary time series-- Denotes a time series whose statistical properties are
independent of time.
Trend pattern-- Gradual shifts or movements to relatively higher or lower values over a longer
period of time.
sh
Th
Seasonal pattern-- Recurring patterns over successive periods of time.
Time series plot not only exhibits a seasonal pattern over a one-year period but also for less than
one year in duration.
This study source was downloaded by 100000831008184 from CourseHero.com on 10-12-2021 14:55:22 GMT -05:00
https://www.coursehero.com/file/38690366/Midterm-Review-for-Descdocx/
Four common accuracy measures
are the mean absolute deviation (MAD), the mean absolute percent error (MAPE), the
mean square error (MSE), and the root mean square error (RMSE).
We focus on MSE( mean square error)
is
ar stu
ed d
vi y re
aC s
o
ou urc
rs e
eH w
er as
o.
co
m
Stationary pattern
1. The moving average technique is probably the easiest extrapolation method for
stationary
data to use and understand.
the moving average technique tends to average out the peaks
and valleys occurring in the original data.
Smaller MSE means more accurate.
Note, however, that the MSE includes and weighs relatively old data with the same
importance as the most recent data. Thus, selecting a forecast based on the total MSE of
the forecasting
functions might not be wise because a forecasting function might have achieved a lower
total MSE by fitting older data points very well while being relatively inaccurate on more
recent data.
Forecasting using moving average technique, assume all the estimation for the future
periods are the same.
2. The weighted moving average technique is a simple variation on the moving average
technique that allows for weights to be assigned to the data being averaged.
3. Exponential smoothing is another averaging technique for stationary data that allows
weights to be assigned to past data.
As shown in the previous equation, the forecast Y^ t11 in exponential smoothing is a
weighted combination of all previous values in the time series where the most recent
observation Yt receives the heaviest weight 1_ 2, the next most recent observation Yt21
receives the next heaviest weight 1_ 11 2 _ 2 2, and so on.
Alpha near 1 means that the forecast reacts faster to the change of data.
Th
Trending pattern
Double exponential smoothing (also known as Holt’s method) is often an effective forecasting
tool for time series data that exhibits a linear trend.
R^2 in regression means that ?% of the variation is accounted for the model
sh
Component of time series:
1. Trend
2. Seasonal
3. Cyclic
Which statement is correct regarding the smoothing coefficient (alpha) in the
exponential smoothing model?
1. A small alpha is best when the series varies substantially
2.
A large alpha is best when the series has little random variability
This study source was downloaded by 100000831008184 from CourseHero.com on 10-12-2021 14:55:22 GMT -05:00
https://www.coursehero.com/file/38690366/Midterm-Review-for-Descdocx/
3. Alpha ranges is value between 0 and 1
What graphic is best for detecting a trend in a time series?
Scatter diagram
The moving average model consists of one parameter namely the number of period in
the moving average.
The purpose of data smoothing is to make the presence of a trend clearer by reducing
the effects of other fluctuations. (look for a trend)
The sum of the digits for three periods is six (1+2+3 =6). Therefore ,the weight for the oldest value is
one-sixth (0.17).
Lesson 3
-
is
ar stu
ed d
vi y re
aC s
o
ou urc
rs e
eH w
er as
o.
co
m
Simple Regression (linear regression)
- Correlation R measures strength, ranges from -1 to 1, very strong when stronger thatn +/0.75
- Residual analysis is also used to determine if the sample data are appropriate for linear
regression, that is, are conditions for regression inference met.
- The residuals should show constant variation and a random distribution around 0.
- Smallest t-test value and largest p-value means this variable has no significant impact on
the function(p value over 5%)
- When correlation between IVs is strong, generally .7 or more (+/-), we would opt to
remove one of those IVs.
- X variable is also called the predictor
- The slope is the same sign as the correlation coefficient
- COD(coefficient of determination, R^2) = (SST-SSE)/SST=SSR/SST
- The method of ordinary least squares is used to fit the regression which is based on the
principle of minimizing the sum or the errors squared.
The error term is symbolized by the Greek letter α.
R^2 ranges from 0 to 1
Basic forecasting ground rules
o Use X value within the database range
o Common sense
o Use only statistically significant variables
Lesson 4
sh
Th
Multiple Regression
- The adjusted R-squared is a modified version of R-squared that has
been adjusted for the number of predictors in the model. The adjusted Rsquaredincreases only if the new term improves the model more than would be
expected by chance. It decreases when a predictor improves the model by less
than expected by chance.
- A dummy variable can take on one of two possible values. Typically these values are zero or
-
one (e.g., male = 0 and female = 1).
The number of dummy variables required to describe a categorical variable is N -1 where N
is the number of categories. A variable that measures four different religions would require
three dummy variables (i.e., 4-1 = 3).
This study source was downloaded by 100000831008184 from CourseHero.com on 10-12-2021 14:55:22 GMT -05:00
https://www.coursehero.com/file/38690366/Midterm-Review-for-Descdocx/
-
-
The adjusted coefficient of determination is the proportion of total variance in the dependent
variable explained by the independent variables adjusted for the sample size and the
number of predictor variables. As the sample size increases the difference between Rsquare and the adjusted R-square decreases.
Discriminant analysis is a variation of multiple linear regression analysis where the
dependent variable is categorical. The simplest model is where Y is binary (e.g., Yes and
No).
Basic regression model assumption
o Linear relationship between X and Y
o Residuals are normally distuributed
o Error terms are independent
-
The F statistic is used to judge the overall statistical significance (performance) of the
multiple regression model. The F statistic is always takes on a positive value. The testing
process is similar to the one used for judging the statistical significance of each f the
independent variables.
A p-value is the probability of obtaining a test statistic (e.g., t statistic) at least as extreme as
the one that was actually observe assuming that the null hypothesis is true. Typically, the null
hypothesis (H0)is rejected if the p-value is less than the stated alpha.
-
is
ar stu
ed d
vi y re
aC s
o
ou urc
rs e
eH w
er as
o.
co
m
-
-
Standard Multiple Regression analysis procedure
o Full
o Backward stepwise
o Forward Stepwise
-
The accuracy of the forecast, as measured by the confidence interval, decreases as the
values of the predictor variables used in the regression forecast varies from their mean
values (both above and below).
eta coefficients (also known as standardized coefficients/slope) are regression model
coefficients that have been standardized resulting in a variance of one. This procedure is
done so that one can identify the relative impact of changes in each predictor variable on the
target variable when each of the predictor variables are measured in different units (e.g.,
dollars, age, gender).
-
Lesson 5
Th
Maximax-large
Maximin-large in min
Minmax- min in large
EMV-large
EOL(expected opportunity loss)-small
EVPI: value of the information, max return of the information
Why is sensitivity important?
Most model coefficients are estimates
sh

Primary problems in decision-making process
-
Conflicting objectives
Uncertainty regarding the future
Lack of data
This study source was downloaded by 100000831008184 from CourseHero.com on 10-12-2021 14:55:22 GMT -05:00
https://www.coursehero.com/file/38690366/Midterm-Review-for-Descdocx/
The probability of exceeding one standard deviation from the mean in a normal distribution is 0.16
Two basic components of payoff matrix: states of nature and alternatives
Implement solutions is not in decision analysis process
EMV: Average monetary outcome of a decision if it was repeated a large number of times
Maximax: non-probabilistic
Lesson 6
sh
Th
is
ar stu
ed d
vi y re
aC s
o
ou urc
rs e
eH w
er as
o.
co
m
Chance nod= event nod
Risk averse: marginal utility of money diminishes
Bayesian: revision of past probabilities
This study source was downloaded by 100000831008184 from CourseHero.com on 10-12-2021 14:55:22 GMT -05:00
https://www.coursehero.com/file/38690366/Midterm-Review-for-Descdocx/
Powered by TCPDF (www.tcpdf.org)
Download