Question 1.
A classification model (in PMML format) was created by a third party vendor tool such as SPSS. 

After importing the model into DB2, where is it stored?

A. In a user-specified table . under an automatically generated name, e.g. SPSS_017.
B. In the table IDMMX.CLASSIFMODELS under a user-specified MODELNAME. The default is the value of the attribute "modelname" in the PMML TreeModel.
C. In the table .CLASSIFMODELS under a user-specified MODELNAME. The default is the value of the attribute "modelname" in the PMML TreeModel.
D. In a user-specified table . under a user-specified MODELNAME. The default is the value of the attribute "modelname" in the PMML TreeModel.

Answer: B

Question 2.
Which of the following must be fulfilled to execute a mining flow that creates a classification model in the database?

A. All input and output ports of all operators must be connected.
B. The flow must contain a Predictor operator whose properties define a Cost Matrix.
C. The flow must contain a Predictor operator whose output port must be connected to a Visualizer.
D. The flow must contain a Predictor operator whose properties (Mining Settings) define a categorical target column.

Answer: D

Question 3.
Application data changes over time and differs from that used to train a classification model M. Assume that a mining flow with a Tester operator with M and the application data as input and a test model T as output has been defined. 

What must be done to check if M can still safely be applied?

A. Open T in a visualizer and check the tree depth. If the tree depth is greater than 10, the model is OK.
B. Open T in a visualizer and check the model quality figures. If the quality is greater than 0.5, the model is OK.
C. Open M and T in a visualizer and check the tree depth. If the tree depth of T is greater than the tree depth of M, the model is OK.
D. Open M and T in a visualizer and check the model quality figures. If the quality is about the same as the quality of M, the model is OK.

Answer: D

Question 4.
As part of a market basket analysis effort, a retail company needs to analyze sales transaction data to understand which products tend to be bought together by the same customer during each visit to the store. 

Which data mining function is most appropriate for this business problem?

A. Clustering
B. Regression
C. Associations
D. Classification

Answer: C

Question 5.
A group of Alphablox users needs a report which contains rule heads and bodies of an Association model in a tabular format. 

Which of the following describes a complete Mining Flow that loads the model output of an Associations operator into a table?

A. Table Source ---> Associations --> Table Target
B. Table Source ---> Associations --> Associations Extractor
C. Table Source ---> Associations --> Visualizer --> Table Target
D. Table Source ---> Associations --> Associations Extractor --> Table Target

Answer: D

Question 6.
An investment banking company wants to use data mining to predict customer lapsing behavior. The cost of losing existing customers is very high, and the company wants to keep all customers happy. Based on a recent customer survey with a small subset of existing customers, 2% of customers indicated they were not satisfied. To predict whether or not a customer is satisfied, Predictor and Tester operators in DWE Design Studio were used to analyze the survey results. 

Which of the following should be done in this situation?

A. Increase the Maximum tolerated systematic error (RSquared).
B. Decrease the Maximum tolerated systematic error (RSquared).
C. Increase the weight of "Satisfied (Actual) & Dissatisfied (Predicted)" in the Cost Matrix  parameter option.
D. Increase the weight of "Dissatisfied (Actual) & Satisfied (Predicted)" in the Cost Matrix parameter option.

Answer: D

Question 7.
An insurance company wants to use data mining to predict fraudulent claims. Based on historical data, 1% of claim transactions were classified as fraud. To work on this problem, the Predictor and Tester mining operators in DWE Design Studio were used to analyze a small subset of claim history data. 

Which preprocessing function operator is the most efficient mechanism to feed data into the Predictor mining operator?

A. Sampler
B. Select List
C. Random Splitter
D. Field Extractor

Answer: C

Question 8.
Which mining model needs to be processed through training and testing phase before being applied into the Scoring function?

A. Sequential
B. Clustering
C. Association
D. Classification

Answer: D

Question 9.
Given the following association rule: 
[Tomato][Carrot]=>[Orange Juice] Support = 10 Confidence = 80 Lift = 4

Which of the following can be deduced from this information?

A. The number of orange juice sales is made up of 20% of all sales transactions.
B. Customers buy the combination of tomato, carrot and orange juice 80% more than expected.
C. There is a 10% chance that customers also buy orange juice when they buy tomato with carrot.
D. Out of 2 million sales transactions, approximately 1,600,000 show that tomato, carrot and orange juice are bought together.

Answer: A

Question 10.
Which of the following mining operator provides a model output that can be directed into the Visualizer operator in the DWE Mining Flow.

A. Associations
B. Cluster Table
C. Predict Column
D. Find Deviations

Answer: A

	Web www.certsbraindumps.com