1

COMP8410 Data Mining 2022

Assignment 2
Maximum marks 100
Weight 25% of the total marks for the course
Length
Maximum of 10 pages excluding cover sheet, bibliography and
appendices.
Layout
A4 margin, at least 11 point type size, use of typeface, margins
and headings consistent with a professional style.
Submission deadline 9:00am, Monday, 9 May
Submission mode Electronic, via Wattle
Estimated time 15 hours
Penalty for lateness 100% after the deadline has passed
First posted: 29th March, 10 AM
Last modified: 29th March, 10 AM
Questions to: Wattle Discussion Forum

This assignment specification may be updated to reflect clarifications and modifications after it is
first issued. It is strongly suggested that you start working on the assignment right away. You can
submit as many times as you like. Only the most recent submission at the due date will be assessed.
In this assignment, you are required to submit a single report in the form of a PDF file. You may also
attach supporting information (appendices) as one or more identified sections at the end of the
same PDF file. Appendices will not be marked but may be treated as supporting information to your
report. Please use a cover sheet at the front that identifies you as author of the work using your u-
number and name and identifies this as your submission for COMP8410 Assignment 2. The cover
sheet and appendices do not contribute to the page limit or word count.
You are expected to write in a style appropriate to a professional report. You may refer to
http://www.anu.edu.au/students/learning-development/writing-assessment/report-writing for
some stylistic advice. You are expected to have both an introduction and a conclusion in your report.
No particular layout is specified, but you should follow use no smaller than 11 point typeface and
stay within the maximum specified page count. Page margins, heading sizes, paragraph breaks and
so forth are not specified but a professional style must be maintained. Text beyond the page limit
will be treated as non-existent.
This is a single-person assignment and should be completed on your own. Make certain you
carefully reference all the material that you use, although the nature of this assignment suggests few
references will be needed. It is unacceptable to cut and paste another author's work and pass it off
2

as your own. Anyone found doing this, from whatever source, will get a mark of zero for the
assignment and, in addition, CECS procedures for plagiarism will apply.
No particular referencing style is required. However, you are expected to reference conventionally,
conveniently, and consistently. References are not included in the page limit. Due to the context in
which this assignment is placed, you may refer to the course notes or course software where
appropriate (e.g. “For this experiment Rattle was used”), without formal reference to original
sources, unless you copy text which always requires a formal reference to the source.
An assessment rubric is provided. The rubric will be used to mark your assignment. You are advised
to use it to supplement your understanding of what is expected for the assignment and to direct
your effort towards the most rewarding parts of the work.
Your assignment submission will be treated confidentially. It will be available to ANU staff involved
in the course for the purposes of marking. It may be shared, de-identified, as an exemplar for other
students.

Task
You are to study the supplied data set and to apply data mining processes and techniques to
discover interesting things about the data. You are to write a short report that justifies and explains
your methods in detail, presents your results, and evaluates and interprets the results you find. In
the following, the task is described in terms of what your report should contain, not in terms of the
steps you should take to carry out the assignment. In your report, similarly, you should describe the
methods used in terms of the language of data mining, not in the terms of commands you typed or
buttons you selected.
1. Introduce the problem
You must provide some context to the data mining project you are working on. You could properly
refer to the purpose of learning and assessment for COMP8410, but in addition you should set some
goals for the exercise – what do you expect to learn from the data? What are you looking for? It is
possible that you may not achieve the goals you set here, but it should be possible to trace the
results you present back to the goals as motivating questions. Furthermore, you should review the
goals you state here in your conclusion.
2. Describe your data
You must
• identify the source of the data and the population over which the data is sampled,
• broadly describe the attributes in the data,
• offer a cursory assessment of data quality, and
• include a basic statistical summary of the data you have.
This should comprise a brief description of the data necessary to explain the context for the work
presented here in a self-contained way, although for more detail it might refer to information
provided with this assignment specification or elsewhere.
3. Describe your methods
You are encouraged to use Rattle or R for this assignment. You may use external tools instead for
part or all of the work (e.g. you might prefer to use python for data pre-processing). Use of
3

alternative tools may make your explanations of methods more wordy, your methods more difficult
to reproduce, and your assignment harder to mark, so take this into account. You will not be
awarded marks for methods where your method cannot be understood.
You must use at least two clearly distinct data mining methods as taught in this course. The distinct
methods should be diverse regarding: i. different problems like classification, prediction, etc. and ii.
different algorithms like NN, DT, etc. You may additionally use multiple other methods taught in this
course. Further, you may choose to use some methods not addressed in this course. You must
justify your choice of methods with reference to the data types involved, the questions you are
looking to answer, the benefit of application to practice, computational feasibility, experimentation
experience, or other reasons.

Application of some methods, or addressing particular questions, may require you to pre-process
the data in some way. For example, if you are looking to predict outcomes independently of time or
political preference, you could consider removing time and political preference attributes from the
dataset. You must include either a statement that no such processing was performed or else brief
information on any
• removal of provided data from consideration,
• imputation or other transformation, or
• differences in the basic data summary from that you prepared from the original
data.
Data pre-processing can be a never-ending task. Be careful to exercise your judgement on how much
you do here, taking account of the marking rubric.

Your description must be sufficient for a reasonably competent professional in the field to
reproduce your major results. You may choose to attach detailed specifications or configuration
parameters as an appendix (which does not contribute to the word count). If you are using methods
that were not taught in the course it would normally be necessary to provide extra detail over that
that can be assumed for methods taught in the course. Extended technical detail may be included in
an appendix or by well-chosen references that contain enough information to implement the
technique.
4. Present your results
You must explain what you found. This should not be a complete listing of everything you found. You
should select results that are interesting, surprising, explanatory, answer your initial questions, or
are otherwise meaningful, and explain why they are meaningful. Your selected results must be
supported by appropriate formal quality measures and must be interpreted within the context of
the problem context you gave. Your interpretation must be pitched towards an expert in the field
related to the data source and business problem but who may not be an expert in data mining. You
might consider using diagrams to assist but use your judgement about any added value of diagrams.
5. Conclude with opportunities for application of your results and identification of further work
Here you should write about the significance of your results and the challenge (or not) of using the
results to make changes in the practice for which your data was collected. This analysis should be
made in the context of the goals you set in your introduction, and you can afford to speculate about
possible impacts of what you found.
You are not expected to be an expert in the area of application, nor to solve challenges you might
raise with putting your results into practice. Identifying further work may include identifying
4

additional data that could be used to refine the results you found, or alternative methods that
should be tried with additional resources.
5

Assessment Rubric
This rubric will be used to mark your assignment. You are advised to use it to supplement your understanding of what is expected for the assignment and to
direct your effort towards the most rewarding parts of the work. Your assignment will be marked out of 100, and marks will be scaled back to contribute to
the defined weighting for assessment of the course.

Review
Criteria
Max
Mark
Exemplary Excellent Good

Acceptable Unsatisfactory
Overall holistic
evaluation of
the report
10 9-10
Highly original and very
interesting.

Excellent, detailed and
relevant discussion that
develops and enhances the
reader's understanding of
the topic.

Very clear key message and
closely associated
conclusion.


7-8
Interesting with some
originality.

Relevant discussion of
sufficient detail to allow the
reader to develop a clear
understanding of the topic.

Identifiable key message and
related conclusion.
6
Interesting but lacking
originality.

Although mostly relevant,
discussion sometimes lacks
sufficient detail to allow the
reader to develop a
consistent understanding of
the topic.

Apparent key message and
associated conclusion.
5
Not very interesting or
original.

Discussion is not always
relevant nor sufficiently
detailed to enable the reader
to develop an understanding
of the topic.

Difficult to be certain what
the key message is and how
the conclusion relates to it

0-4
Boring and mundane.

Discussion lacks detail, is
mostly irrelevant and doesn't
help the reader to develop
an understanding of the
topic.

No discernible key message
or conclusion.

Communication,
Structure and
Presentation
10 8-10
Exemplary use of language
enhancing the quality of the
submission.

Very well ordered with
logical and clear structure
supported by appropriate
headings and sub headings.

All use of others' ideas and
materials acknowledged.
References are all included
and are formatted
7
Very good use of language.

Well-ordered and logical.
Headings and sub-headings
help to clarify text.

All use of others' ideas and
material is acknowledged.
All references are included,
though some minor
inconsistency of in-text
citation or formatting.

6
Reasonable but needs some
revision.

Mostly well-ordered and
logical, most supported by
headings and sub-headings

All use of others' ideas and
material is acknowledged.
Some references are missing
and occasional
inconsistencies of in-text
citation and formatting.
5
Poor, needs significant
revision.

Order is not always logical
and is sometimes confusing.
Headings are largely those
suggested by the assignment
specification and the
questions posed.

All use of other's ideas and
material is acknowledged,
though sometimes
0-4
Very difficult to understand.

Order is confusing and not
always logical. Headings and
sub-headings do little to help
clarify the text

Not all use of other's ideas
and material is
acknowledged. Missing in-
text citations, i.e. plagiarism.
References in the
bibliography not used in the
6

Review
Criteria
Max
Mark
Exemplary Excellent Good

Acceptable Unsatisfactory
consistently and
appropriately.

Diagrams and/or images are
ideally suited to the points
where they are used.
Diagrams and/or images are
used effectively.

Diagrams and/or images
improve readability.
inconsistently. Missing
references and inconsistent
in-text citation and
formatting.

Diagrams and/or images are
not well selected.
text. Poorly and
inconsistently formatted.

Diagrams and/or images
detract from the key
messages.
Problem
Description

10 9-10
Goals are clear, challenging,
and suitable for the data
used.

Wider context of goals is
discussed (e.g. expected
impact or importance).

The problem description
provides context for the
data mining that is
connected and used in the
rest of the work.

7-8
Problem description is clear
and suitable for the data
used.

The problem description
provides context for the
data mining that is
connected and used in the
rest of the work.



6
The problem description
provides adequate context
for the mining work
although some key elements
could be expanded to
support richer analytical
work.

5
Problem description is barely
adequate for the purpose.

Problem description does
not connect tightly with the
work performed.
0-4
Key elements of the problem
description are missing or
insufficiently explained.
Data
Description
10 9-10
Source, attributes,
population, quality
assessment and basic
statistical summary
provided.

Description demonstrates
deep understanding of the
data.
7-8
Source, attributes,
population, quality, and
basic statistical summary
provided.

Data interpreted in terms of
challenges for mining or
suitability for the problem
goals.





6
Most of the source,
attributes, population,
quality, and basic statistical
summary provided.


5

Some of the required
information provided and
correct.
0-4

Required information not
provided and/or incorrect or
misleading, demonstrating
lack of engagement with the
data.
7

Review
Criteria
Max
Mark
Exemplary Excellent Good

Acceptable Unsatisfactory
Method
description
30 24-30
At least 2 course methods
applied, plus at least one
more method which may or
may not be taught in the
course.

R, Rattle and other tools
have been properly
identified and used
appropriately.

Reproduction of major
results is possible from
description of methods.
Data pre-processing is well-
suited to the methods used
and the mining goals, with
justification (or no-pre-
processing, with convincing
justification).
Careful parameter setting
and tuning explained and
justified by experimentation
or theory or both.
Justification for methods
chosen demonstrates careful
attention to the applicability
and limitations of the
methods to the problem
goals.
21-23
At least 2 course methods
applied.

R, Rattle and other tools
have been properly
identified and used
appropriately.

Reproduction of major
results is possible from
description of methods.

Extensive, directed,
experimentation with data
preprocessing or tuning
parameters explained.

Justification for methods
chosen is clear and linked to
problem goals.

18-20
At least 2 course methods
applied.

R, Rattle and other tools
have been properly
identified and used
appropriately.

Reproduction of major
results is possible from
description of methods.

Some experimentation with
data preprocessing or tuning
parameters evident.

Justification for methods
chosen is clear.


15-17
At least 2 course methods
applied.

Not always clear what
software tools were used.
Unclear that reproduction of
results is possible from
description of methods.

Experimentation with data
pre-processing or tuning
parameters barely evident,
suggesting a simplistic
approach to the problem.

Weak justification for
methods chosen.



0-14
Less than 2 course methods
applied.

Not clear what software
tools were used

Methods not described in
adequate detail for
reproduction.


Justification for methods
chosen absent or
unconvincing.
8

Review
Criteria
Max
Mark
Exemplary Excellent Good

Acceptable Unsatisfactory


Results 20 17-20
Outstandingly useful and
potentially actionable results
found.

Results are clearly
interpreted for a domain
expert.

Results supported by well-
selected quality measures,
explained in terms of impact
for domain expert.
14-16
Results presented are well
selected for significance to
the mining goals.

Results are interpreted for
domain expert.

Results are well supported
by selected quality measures
that are explained for
domain expert.
12-13
Major results are clearly
presented, with quality
measures present but
overall interpretation for
domain experts could be
sharper.


10-11
Results are clearly
presented, with typical
quality measures present.

0-9
Scant attention to evaluation
appropriate to the methods
used.

Conclusion and
further work
10 9-10
Thoughtful analysis of how
results could be applied,
including identifying
challenges.

Ideas for further work are
creative, relevant and
exciting, and tied to
application context.


7-8
Analysis of potential
application of results ties to
problem goals and
recognises the application
context.

Ideas for further work are
significant and realistic.
6
Statement of how results
could be applied to the
domain is realistic in the
context of the problem
goals.

Ideas for further work are
present but could be better
tied to the problem and
application domain.
5
Statement of how results
could be applied to the
domain is given.

Possible further work
identified.
0-4
Missing analysis of
application or extension.