Data Analytics: Exploiting the Data Warehouse - Data Warehouse Systems: Design and Implementation - page 375

warehouse to build the data set. Add a Boolean attribute Buyer to
each record: if a customer has purchased an item of (sub)category
X , classify the record as 1, otherwise as 0. Add also an attribute
age.
(b) Analyze the model using Analysis Services. Comment on the main
characteristics, the attributes that the algorithm has selected to
partition the tree, and other features you consider of interest.
Modify the parameters and verify if you obtain the same model.
(c) Write a prediction query that checks for a new customer if she will
order an item of the required (sub)category.

9.3 Consider again the Foodmart data warehouse:

(a) Use the Customer dimension and build a clustering model.
(b) Analyze the model using Analysis Services. Comment on the main
characteristics, the number of clusters, and so on. Analyze if the
partition reflects correctly the data or if you think that the model
must be revised. Modify the parameters to produce other models.
(c) Write a DMX query that returns the cluster to which a new
customer will most likely belong. Assume that this customer is a
35-year-old single female with university studies.

9.4 Consider the following transaction database:

TID Items

{
{

}
}

T1 A,B,C

T2 A,B,D

{
{
{
{
{
{
{

}

T3 B,C

}
}
}
}

T4 D,E,F

T5 E,F,G

T6 A,C,E

T7 A,B,D

}
}

T8 A,B,C,F

T9 A,D,E,F

T10 { B,C,D,E }

(a) Manually run the Apriori algorithm to find out the frequent
itemsets and rules with minimum support and confidence of 40%.
(b) Use the FUP algorithm to insert the following transactions:

TID Items
T1

{}
{ }
{}
{}

A,K

T2 C,E,K

T3 F, G

T4 K,L