Exam 17: Business Analytics

SCENARIO 17-7 The output below shows the results of cluster analysis on the different regions of the world (North America, Asia Pacific, Western Europe, Central & Eastern Europe, Latin America, and Middle East & Africa)based on the amounts (in billions of US dollars)and the percentage changes from prior year of business-to-consumer e-commerce sales last year. -Referring to Scenario 17-7, North America and Western Europe are the second group of countries that cluster together.

SCENARIO 17-3 The tree diagram below shows the results of the classification tree model that has been constructed to predict the probability of a cable company's customers who will switch ("Yes" or "No")into its bundled program offering based on the price ($30, $40, $50, $60)and whether the customer spends more than 5 hours a day watching TV ("Yes" or "No")using the data set of 100 customers collected from a survey. -Referring to Scenario 17-3, the highest probability of switching is predicted to occur among customers who do not watch more than 5 hours of TV a day and are offered the bundled price of lower than $50.

Which of the following is NOT among the predictive analytics methods covered in the book?

Most information design specialists prefer bullet graphs over gauges because bullet graphs foster the direct comparison of each measurement.

Neural networks do not make any a priori assumption about the distribution of the data and, hence, are nonparametric methods.

SCENARIO 17-1 The table below contains the sparklines for the rates of return (in percentage)for three different stocks from 2007 to 2013. -Referring to Scenario 17-1, the sparklines enable you to conclude that the rates of return of the stock market in general are volatile from 2007 to 2013.

In a classification tree, the dependent variable is a categorical variable.

SCENARIO 17-4 The regression tree below was obtained for predicting the weekend box office revenue of a newly released movie (in thousands of dollars)based on data collected in different cities on the expenditure (at $25, $30, $35, $40, $45, $50, $55, $60, $65 or $70 thousand)spent on TV advertising and the number of times (10, 15, 20, 25, 30 or 35)a day the advertisement appear on TV. -Referring to Scenario 17-4, the first split occurs at $45 thousand spent on TV advertising.

The result of the regression tree is affected by the distribution of the independent variables.

Some business analytics involve starting with many variables, followed by filtering the data by exploring specific combinations of categorical values or numerical range.In Excel, this approach is mimicked by using a slicer.

Some business analytics involve starting with many variables, followed by filtering the data by exploring specific combinations of categorical values or numerical range.In Excel, this approach is mimicked by using gauges.

Some consider bullet graphs little more than examples of chartjunk, even as many decision makers request them due to their visual appeal, due to the amount of the space they consume.

Which of the following finds relationships in data that may not be readily apparent?

SCENARIO 17-4 The regression tree below was obtained for predicting the weekend box office revenue of a newly released movie (in thousands of dollars)based on data collected in different cities on the expenditure (at $25, $30, $35, $40, $45, $50, $55, $60, $65 or $70 thousand)spent on TV advertising and the number of times (10, 15, 20, 25, 30 or 35)a day the advertisement appear on TV. -Referring to Scenario 17-4, the highest mean weekend box office revenue is predicted to occur with $55 thousands spent on TV advertisement and 35 advertisement appearances a day.

Some business analytics involve starting with many variables, followed by filtering the data by exploring specific combinations of categorical values or numerical range.

SCENARIO 17-7 The output below shows the results of cluster analysis on the different regions of the world (North America, Asia Pacific, Western Europe, Central & Eastern Europe, Latin America, and Middle East & Africa)based on the amounts (in billions of US dollars)and the percentage changes from prior year of business-to-consumer e-commerce sales last year. -Referring to Scenario 17-7, at the two-cluster level, Asia Pacific alone makes up one of the two clusters.

SCENARIO 17-3 The tree diagram below shows the results of the classification tree model that has been constructed to predict the probability of a cable company's customers who will switch ("Yes" or "No")into its bundled program offering based on the price ($30, $40, $50, $60)and whether the customer spends more than 5 hours a day watching TV ("Yes" or "No")using the data set of 100 customers collected from a survey. -Referring to Scenario 17-3, what percentage of the variation in whether a customer will switch into its bundled program offering can be explained by the price and whether the customer spends more than 5 hours a day watching TV?

In multidimensional scaling, the general rule in to increase the number of dimensions as long as the stress statistic increases substantially.

Each observation is treated as its own cluster at the beginning in k-means clustering.

The LogWorth statistic is a measure of the probability that can be attributed to the response that has occurred.