Database Reference
In-Depth Information
then proceeds on topics about data exploration and pre-processing, followed by basic
mining and modelling techniques and evaluation of patterns. The basic techniques are
then followed by advanced and alternative techniques. The module covers application
issues towards the end. This approach of module delivery enables the students to gain
knowledge from classes and then apply it practically in the project.
2.5 Assessment Framework of the Project
The project is assessed according to the quality of work at each stage of the data min-
ing process. Correctness, completeness and soundness of judgement are the main
factors for determining marks. Table 1 presents a framework for project assessment.
The matrix highlights the assessment focuses by each factor at each stage. Further
details for assessment can be found in [2].
Table 1.
Assessment Framework
Phases
Data
Understanding
& Exploration
(20%)
Data
Preparation &
Pre-processing
(25%)
Data Modelling
& Mining of
Patterns (25%)
Evaluation of
Result Patterns
(20%)
Factors
Correct mining
tasks, sensible
choices of
modelling
solutions.
Sensible setting
of parameters.
Correct
understanding of
evaluation
metrics and
interpretation of
patterns
Correct
understanding
of data
characteristics
and features
Correct data
pre-processing
and preparation
operations
Correctness
Complete
collection,
summarisation
and
categorisation of
patterns.
Evaluation of
both pros and
cons.
Using
alternative
solutions.
Alternative
setting of
parameters.
Comparison of
solutions.
Coverage of
aspects of data
features such as
data types,
distributions,
missing values,
etc.
Sufficiency of
the operations
for the purpose
of discovery
Completeness
Justification for
selection of
mining
operations and
parameter
settings
Need for further
mining.
Selec-
tion/Identificatio
n of interesting
patterns.
Justification for
the operations
and their
relevance to
mining
Soundness of
Judgement
Needs for data
pre-processing
Project Planning, Execution, Management and Teamwork (10%)
Given the importance of all the stages, marks should be evenly divided. Because of
the complexity and amount of time required, the Data Preparation and Pre-processing
stage and the Data Modelling/Mining stage may take a marginally larger share of the
total mark than the other stages. A certain small percentage of the total mark may be
given to the successful planning, effective execution and management of the project,
