The manager of the commercial loan department for a bank wants to develop a rule to use in determining whether or not to approve various requests for loans. The manager believes that three key characteristics of a company's performance are important in making this decision: liquidity, profitability, and activity. The manager measures liquidity as the ratio of current assets to current liabilities. Profitability is measured as the ratio of net profit to sales. Activity is measured as the ratio of sales to fixed assets. The manager has collected the data found in the file Loans.xlsm accompanying this book containing a sample of 98 loans that the bank has made in the past five years. These loans have been classified into two groups: (1) those that were acceptable and (2) those that should have been rejected.

a. What are the coordinates of the centroids for the acceptable loans and the unacceptable loans? Round your answers to two decimal places, if necessary.

	Centroid
Group	Liquidity		Profitability		Activity
1
2

b. Use XLMiner's standard data partition command to partition the data into a training set (with 60% of the observations) and validation set (with 40% of the observations) using the default seed of 12345. Use discriminant analysis to create a classifier for this data. How accurate is this procedure on the training and validation data sets? Round your answers to one decimal place.

Data Set	Overall Error
Training	% %
Validation

c. Use logistic regression to create a classifier for this data. How accurate is this procedure on the training and validation data sets? Round your answers to one decimal place.

Data Set	Overall Error
Training	% %
Validation

d. Use the k-nearest neighbor technique to create a classifier for this data (with normalized inputs). What value of k seems to work best?

How accurate is this procedure on the training and validation data sets? Round your answers to one decimal place and if your answer is zero, enter "0".

Data Set	Overall Error
Training	% %
Validation

e. Use a classification tree to create a classifier for this data (with normalized inputs and at least 4 observations per terminal node). Create a graphic depiction of the best pruned tree using the validation data. How accurate is this procedure on the training and validation data sets? Round your answers to one decimal place.

Data Set	Overall Error
Training	% %
Validation

f. Use a manual neural network to create a classifier for this data (rescale the data using standardization, use a single hidden layer with 3 neurons, use a stopping rule on training only with 300 epochs and a maximum of 50 epochs without improvement). How accurate is this procedure on the training and validation data sets? Round your answers to one decimal place.

Data Set	Overall Error
Training	% %
Validation

g. Return to the Data sheet and use the Transform, Bin Continuous Data command to create binned variables for liquidity, profitability, and activity. Use XLMiner's standard data partition command to partition the data into a training set (with 60% of the observations) and validation set (with 40% of the observations) using the default seed of 12345. Now use the nave Bayes technique to create a classifier for the data using the new binned variables for liquidity, profitability, and activity. How accurate is this procedure on the training and validation data sets? Round your answers to one decimal place.

Data Set	Overall Error
Training	% %
Validation

h. Which of the classification techniques would you recommend the company actually use? SelectDiscriminant analysisLogistic regressionk-nearest neighbork-nearest neighbor (k optimized)Classification treeNeural networkNaive BayesItem 20

i. Suppose that the manager receives loan applications from companies with the following financial information. According to your recommended classifier, which of these companies do you expect to be acceptable credit risks?

Company	Loan
A	Acceptable or Not acceptable?
B	Acceptable or Not acceptable?
C	Acceptable or Not acceptable?
D	Acceptable or Not acceptable?
E	Acceptable or Not acceptable?