User interface language: English | Español

Date November 2017 Marks available 3 Reference code 17N.2.HL.TZ0.4
Level HL Paper 2 Time zone no time zone
Command term Distinguish Question number 4 Adapted from N/A

Question

A telecommunication company stores a large amount of data in three databases.

The database segmentation is carried out on the CUSTOMERS database.

Data mining is used to extract knowledge hidden in this large amount of data. Before using data mining processes the existing data should be cleaned up.

Customers committing fraud is a risk to the company.

Evaluate the use of an object-oriented database as opposed to a relational database.

[4]
a.

Define a spatial database.

[2]
b.

State what is meant by database segmentation.

[1]
c.i.

Explain one benefit of database segmentation to the telecommunication company.

[2]
c.ii.

Explain how ETL processes could be used in data preparation.

[3]
d.

Distinguish between the use of association and sequential patterns as data mining techniques.

[3]
e.

Describe how deviation detection could be used to detect fraud at the telecommunications company.

[2]
f.

Customers who decide to leave the telecommunication company for a competitor may result in huge losses for the telecommunications company.

Explain with the use of an example, how predictive modelling could be used to optimize information sent to existing customers.

[3]
g.

Markscheme

Award up to [4 max].
Object-oriented databases store objects rather than data such as integers, strings or real numbers;
Objects are used in object-oriented languages such as C++, Java, and others;
Objects consist of attributes (data which defines the characteristics of an object) and methods (procedures or functions which define the behaviour of an object) / the data is closely related to the programs that process them;
Which may be a more accurate representation of the real-world objects they represent;

Relational databases store data such as integers, strings or real numbers in tables / uses tables with rows and columns);
Relational database is normalized so data is not repeated more often than necessary;
All columns depend on a primary key (a unique value in the column) to identify the column;
Once the specific column is identified, data from one or more rows associated with that column may be obtained or changed;
There is a distinct separation between the storage of data and any application using them;

Disadvantages of object-oriented database over relational database
Lower efficiency when data is simple and relationships are simple;
Relational tables are simpler;
More user tools exist for relational databases;
Standards and support for relational databases are more stable so changes in database are less likely to be required;

a.

Award [2 max].
A spatial database stores:
Spatial data / locational data / geometric data types;
Such as point, polygons, lines, 3D shapes, coordinates etc.; // allow any one of these;
That model the structure of geometric/3D objects;
And make use of spatial indexes to access data;
And supports geometric functions;

b.

Database segmentation means dividing entities into groups with similar characteristics/common attributes;

c.i.

Award [1] for identifying a benefit and [1] for a further explanation, up to [2 max].

Benefit
Increased profit;
Better reputation;
Increased number of customers;
Better opportunities for growth;
To gain a larger share of the market;

Explanation
By studying the groups and behaviour of the customers in the group (for example, average call time of customers younger than 30 or which services are purchased by customers older than 60, etc.);
The company will better understand customers and serve them better;
Can better decide which marketing actions should be taken;
Can better predict which group of customers new products/services should be offered to;
Can better decide which new products/services are needed;

c.ii.

Award up to [3 max].
Useful data should be discovered then extracted from various sources;
Transformations performed;
Repair inaccuracy between data formats;
Remove mistakes/spelling mistakes;
Delete unnecessary data/data fields / clean up the data;
(accept other examples)
Load data into data warehouse;

d.

Award up to [3 max].
Association is a (rule based) data mining technique which searches for relationships between data items;
In the same transaction;
Sequential patterns is a technique that seeks to discover/identify similar patterns in transaction data;
Over a period of time;

Note: Accept examples in given situation.
Association could find out that customers who buy a new mobile phone also change the subscription type;
Sequential patterns could discover that twice a year number and duration of calls increase (e.g., customers’ birthday and New Year’s Eve);

e.

Award [1] for creating a model and [1] for detecting changes/differences in the data/model from previously measured values, up to [2 max].

Example answer
The customer call records and calling behaviour (other activities) could be summarized to obtain their calling pattern;
The (previous) calling pattern could be compared with the customers’ recent call/activity;
In case of significant changes/differences the activity is suspicious/fraud detected;

f.

Award [1] for creating a statistical model, [1] for analysing the model and forecasting probabilities/trends and [1] for an appropriate example in the given situation.

Example answer 1
Customer’s online behaviour pattern could be created and considered;
Customers who spend less time logged on may be less likely to renew their annual subscription;
In this case, the ability to keep the customer can be increased by offering new/better/cheaper services and products;

Example answer 2
Customer’s calling pattern could be created and analysed;
If, for example, the customer has high monthly usage of services then the customer will probably renew his annual subscription;
(No actions made by the company needed);

If a customer is younger than 20, may spend more time exploring new service features, and the customer is at a greater risk to cancel the contract/the customer will probably not renew annual subscription;
(Actions made by the company might be needed, for example: offering new/better services and products;)

g.

Examiners report

[N/A]
a.
[N/A]
b.
[N/A]
c.i.
[N/A]
c.ii.
[N/A]
d.
[N/A]
e.
[N/A]
f.
[N/A]
g.

Syllabus sections

Option A: Databases » A.4 Further database models and database analysis
Show 33 related questions
Option A: Databases

View options