CISB Final

The Excel formula used to determine correlation between two variables is:

=CORREL( )

In exponential smoothing, the weight given to the actual value (not the forecast) in period t is 1-α

False

In exponential smoothing, the weight given to the actual value in period t is called the smoothing consideration.

False

In order to get the best possible regression model, we want to make sure the errors are as large as possible.

False

Moving average and exponential smoothing are time series forecasting methods appropriate with data that shows a sharp trend, cycle or seasonal effect.

False

The following graph of residuals represents a linear relationship between independent variables.

False

The goal of cluster analysis is to organize observations into dissimilar groups.

False

The naïve forecasting method is the same as a moving average for k=3.

False

Which type of time series pattern is shown below?

Seasonal

If a marketing company finds that sales increase for employees for the first few months at the job, but then find after 8 months burn out sets in and sales begin to decline, then which of the following is true?

The regression model is a piecewise model and the knot is about 8 months.

Which of the following accounts for the variability in y that cannot be explained by the linear effect on the q independent variables?

ε

In a linear regression, which of the following does not apply to the dependent variable?

It is the variable being used to predict the value of the dependent variable

What does sum of squares due to error (SSE) measure?

SSE measures the error from using the estimated regression equation to predict the values of the dependent variable in the sample

In the Venn diagram below, there are two events A and B. What does the rectangle represent?

Sample space

If we run our inference test and find that we do not reject the null hypothesis that beta, β = 0, then what does this mean?

That there isn't a relationship between that particular independent variable and the dependent variable

A hierarchal clustering approach begins with each observation in its own cluster and then iteratively combines the two clusters that are the least dissimilar (most similar) into a single cluster.

True

A piecewise linear regression fits two linear regressions that are joined at some value in which the relationship between the dependent and independent variables change.

True

A simple linear regression model is first specified with population parameters and then when you get data, you can find the estimated simple linear regression equation.

True

A word cloud on tweets can provide insight on trending topics.

True

Averages of past values can provide for a more accurate forecast then the most recent data observation.

True

Before running a forecast method, it is a good idea to run descriptive statistics and graph your data so that you can see what type of pattern the data has (trend, horizontal, seasonal, cyclical).

True

One concern with using past data to predict the future is if the _____________

Historical patterns do not continue into the future.

In the residual scatter below, do the residuals violate any assumptions for valid inference testing?

No

In the below estimated regression line, what is the the slope of the estimated regression line? (Answers are rounded.)

-2.3

In the below decile-wise lift chart, if there are 1,000 customers in total and 50 out of 1000 in the entire sample are expected to default, what is the count of customers in the first decile that actually default, for the first decile given that the height of the bar that corresponds to the first decile is 2.0 (the decile mean/global mean)?

20

If a vending machine company believes the sales territory (North, West, East or West) is an important predictor to total sales (in dollars), how many dummy independent variables should be added to the model?

3

What is the Euclidian distance between the two observations (customers)? u = (2, 5) v = (4, 8)

3.61

If we know that the probability of A is 0.2 (it rains today) and we know the probability of the intersection of A (it rains today) and B (the electricity goes out) is 0.6, then what is the probability of B given A?

30%

A study found that 25% of people like Coke only, 30% of people like Pepsi only and 15% of people like Coke and Pepsi. What is the probability that a consumer likes Coke only, Pepsi only, or both?

40

If we found that r squared = 0.45 in the Butler Trucking example (see textbook), then which of the following does not apply?

45% of the variability in the values of miles traveled in our sample can be explained by the linear relationship between miles traveled and travel time

A company found in a survey that 40% employees left the company due to salary concerns, 20% left due to work concerns, and 10% left due to both -- salary and work -- concerns. Which of the following is true?

50% of employees left for salary or work concerns

If a sales manager finds that 10% of her in-store visitors resulted in actual sales, what is the probability that a customer comes into the store and doesn't buy anything?

90 percent

What does the coefficient of determination (r squared) measure?

A goodness of fit of the estimated regression line

Jaccard's coefficient is defined as follo

A method to measure similarity between observations that doesn't include observations where both observations are "0".

A random variable is...

A numerical description of the outcome of an experiment.

Examples of unstructured data used in text mining:

All of the above

For an association rule to be useful...

All of the above.

What is the difference between an outcome and an event?

An outcome is heads or tails when you flip a coin; Event is a collection of outcomes, heads and tails.

Which probability distribution is often characterized as having two outcomes, Success (p) and Failure (1-p)?

Binomial Probability Distribution

If a random variable is any value within a given range then it is a ____________random variable.

Continuous

The following table gives examples of discrete or continuous random variables?

Continuous

Multicollinearity refers to the _____________between independent variables.

Correlation

Which kind of time series pattern is most pronounced in the graph below?

Cyclical

In a multiple regression model, what do the betas (β) represent?. The change in the mean of the _____________ variable y that corresponds to a one-unit increase in the ____________________ variable, x, holding the values of all other _________________variables in the model constant.

Dependent, independent, independent

Flipping a coin is an example of a....

Discrete random variable

An estimated linear regression produces b0 and b1. b1 is the ..

Estimated change in the mean of the dependent variable y that is associated with a one-unit increase in the independent variable x

The objective of time series analysis is to uncover a pattern in the time series and then ______________________ the pattern into the future.

Extrapolate

A casual model is one in which the independent variables such as years or quarters.

False

A random variable that can take on only specified discrete values is referred to as a continuous random variable.

False

A word cloud is a visual presentation of a document in which the size of the word is proportional to its conditional probability.

False

Cross-validation is an important step in the handout method.

False

Exponential smoothing uses a simple average of past time series values as a forecast .

False

Which of the following is NOT a recommended variable selection procedure?

Forward elimination

The antecedent refers to the ________ portion of the association rule.

If

If you do find that you do not reject that beta = 0 for one particular independent variable, how do you handle the respective independent variable associated with that beta?

If that particular variable doesn't belong theoretically and when you drop the variable it actually improves the model fit (r squared), then drop it.

If you find that in your model that the relationship between the dependent variable and one independent variable is different at various values of a second independent variable, you have what kind of effect in your model?

Interaction

Which of the following is NOT a time series pattern?

Inflationary

The overall error rate in Performance Measures calculates an aggregate measure of miscalculation, but its limitation is...

It does not classify a false positive the same as a false negative.

Which is better, a low p-value or high p-value?

Lower the better

Which of the following is NOT a categorical variable?

Miles traveled

Which of the following is NOT an agglomeration method (method to measure dissimilarity between 2 clusters).

Multiple linkage

What type of probability distribution is characterized here? A company wants to know the probability that its vacuums last more than 1,000 hours.

Normal

Which probability distribution is characterized here? A oil/gas company keeps track of the number of leaks in 100 miles of pipeline.

Poisson

Given the following Confusion Matrix, among those that did not default, how many were misclassified?

Predicted ClassActual Class0 (nondefault)1 (default)0 (nondefault)n00 = 1000n01 = 7501 (default)n10 = 50n11 = 70 50

How is conditional probability defined?

Probability of one event occurring given that some other related event has already occurred

Text mining refers to...

Process of extracting useful information from unstructured data.

A hypothesis is ________________ when the corresponding p value is smaller than some predetermined level of significance, usually 0.05 or 0.01.

Rejected

Explain in words what the below estimated linear regression is measuring. Source: Britannica

Stress test score is the independent variable, being used to explain the variation in the blood pressure reading, the dependent variable

What is the difference between the Euclidian distance and the Jaccard's Coefficient distance measures between observations?

The Euclidian is used for quantitative numbers; the Jaccard's is used to calculate distance between categorical data.

In the below graph, what does e represent?

The difference between the actual observed data for any value and the estimated/predicted data value for that same data point

Why is the lift ratio preferred to the confidence of a rule?

The lift ratio uses conditional probability. For example, it first finds the probability that a customer buys bread and jelly and then finds the probability that a customer also bought peanut butter.

In a linear regression, which of the following defines the independent variable?

The variable used to predict the value of the dependent variable

The consequent refers to the ________ portion of the association rule.

Then

In unsupervised learning...

There is no outcome variable to predict.

Which of the following is NOT a condition necessary for valid inference in a least squares regression model?

There must be at least three independent variables in the model

Why might a NASA scientist use probability theory?

To help managers select the option that has the best chance of success.

Which kind of time series pattern is most pounced in the below graph?

Trend

Which of the following is NOT a discrete probability distribution?

Triangular Probability Distribution

In Chapter 7, inference testing refers to the process of determining whether there is a statistically significant relationship between the dependent variable and each independent variable.

True

Larger values of k -- rather than smaller -- in a moving average approach to forecasting will be able to smooth out random fluctuations in the data.

True

Moving average and exponential smoothing are time series forecasting methods appropriate with data that shows a horizontal pattern.

True

Probabilities are used to determine the chance of success or failure of some event occurring.

True

Probability is a means to understanding and measuring uncertainty. Uncertainty is a factor in virtually all business decisions, so an understanding of probability is essential to modeling such decisions and improving the decision-making process.

True

Regression analysis can be used to forecast data that has a trend, or seasonal pattern, or both.

True

The objective of forecast accuracy is to determine a good quantitative forecasting model.

True

The reason we calculate forecast error, MFE, MAE, MSE and MAPE is to be able to compare forecast methods (naïve, moving average, exponential smoothing) and then select the forecast method that minimizes forecast errors.

True

There are two types of cluster analysis studied in this chapter, k-means clustering and hierarchical clustering.

True

To avoid overfitting your model, you can use only independent variables that you expect to have real/meaningful relationships with the dependent variable; don't let software dictate your model; and use your own judgement to refine your model.

True

When a "trend" variable is added to a regression analysis, it is a series of values, 1, 2, 3, 4, etc. representing time (or it can be years, or quarters, or weeks, etc.).

True

What is another name for descriptive data-mining methods?

Unsupervised learning

Cluster analysis is....

Used in marketing in a process known as market segmentation.

Which of the following does not apply to association rules?

Used to help forecast out-of-sample sales.

If 20 pairs of jeans were sold by Jean's Jeans, and Jean forecasts that 20 pair will sell next week, this is an example of naïve forecasting which is defined as follows:

Using the most recent data to predict future data.

Which of the following does NOT describe a quadratic regression model?

When you observe a linear relationship between sales (dependent variable) and numbers of years employed (independent variable).

Does the following residual scatter plot violate any of the conditions necessary for valid inference?

Yes

Quantitative forecasting methods can be used in all of the following cases EXCEPT:

You are involving the use of expert judgment.

What is of the following is NOT a problem created from overfitting?

You get too many independent variables in the model such that interpretation of the independent variable on the dependent variables gets masked by all the relationships.

If you have three dummy variables in a regression model (as below) where x1 = Region A, x2 = Region B, x3 = Region C, what is the interpretation of b1? Assume y hat = predicted sales.

b1 is the estimated difference between the mean number of sales between Region A and Region C.