The following data is a sample from a loan history database of a Japanese bank

Clients: 1, 2, ..., 10

Clients with approved loan: 1, 2, 6, 7, 8, 9

Clients with rejected loan: 3, 4, 5, 10

Client data:

Unemployed clients: 3, 10

Not married: 1, 2, 5, 6, 7

Amount of money in a bank (Client ID=Money x 10000 yen): 1=20, 2=10, 3=5, 4=5, 5=5, 6=10, 7=10, 8=15, 9=20, 10=5

Using the above information do the following:

1. Chose proper values for the attributes (e.g. yes/no or true/false for nominal attributes and numbers for numeric attributes) and fill in the following relational table:

ID	Employed	Married	Money	Approved
1
2
3
4
5
6
7
8
9
10

2. Create and draw a concept hierarchy for the Money attribute as explained in Lecture notes Chapter 2 (e.g. by grouping the numeric values in 2-3 sets or intervals). Then create another relational table representing the data (as the one in Problem #1) with nominal values for the Money attribute taken from the top level of the concept hierarchy.

3. Using the relational table from Problem #2 create a complete data cube (represented as a relational table with ALL values) by aggregating the values of Approve. Do you need the ID attribute for this purpose? Also, note that some combinations of attribute values may not be present in data and we don't know the value of Approved for these combinations. How do we aggregate them? A common approach used in ML to deal with this situation is the so called Closed World Assumption (CWA), where we assume one of the two possible values. You may use CWA or another approach. Add some comment on this.

4. Create a two-dimensional data cube (represented as a table) using the dimensions Employed and Money (nominal), aggregating the values of Approved.

5. Drill-down the cube from Problem #4 over the Money dimension using the hierarchy built in Problem #2 (climb down the hierarchy to the level of numeric values).

6. By observing the data cubes from Problems #4 and #5 can you identify any trends, regularities or patterns in the data?