QBUS6850_Project_Marking_Criteria
QBUS6850
Group
Project
–
Marking
Criteria
Part
Criteria
Max
Marks
A
Data
Pre-‐processing:
• With
the
best
effort
to
clean
up
the
data,
by
checking/deleting
incomplete
information,
missing
values
and
their
percentages
• Argument
on
not
applying
pre-‐processing
procedure
8
B
EDA:
• Coherent
and
thorough
EDA
with
sufficient
explanation
• Good
argument
and
rationale
for
your
EDA
• Excellent
Presentation
of
Analysis
Results
20
C
Benchmark
Model:
• Successfully
build
a
logistic
regression
model
with
relevant
features
of
training
and
validation
datasets
• A
demonstrated
validation
process
with
your
analysis
and
careful
choice
of
hyperparameters
(if
any)
• Well
documentation
of
the
findings
and
justification
15
D
Feature
Selection
• Successful
implementation
of
feature
selection
and
description
of
chosen
features
• Sufficient
explanation
of
your
choice
and
justification
• Full
report
of
model
setting
and
comparisons
20
E
Use
of
Extra
Features:
• Sufficient
test
and
clear
report
of
using
extra
feature
(even
this
is
not
part
of
your
final
competition
model)
• Comparison,
analysis
and
final
conclusion
of
using
extra
features
15
F
Presentation:
• Clear
presentation
of
text,
figure
and
formulae,
free
of
spelling
and
grammar
mistakes,
etc
7
G
Competition:
• Correct
name
and
format
for
the
final
result
csv
file
• Ratio
compared
to
the
best
performance
among
all
the
groups
15
Applicable
Penalty
Rules:
• Late
submission
penalty
according
the
school/university
policy
• Failure
to
follow
submission
instruction
such
as
not
using
required
file
names:
up
to
10
marks
• Missing
code
submission:
up
to
10
marks
[this
refers
to
the
separated
code
submission!]
• Missing
meeting
minutes,
or
without
details
and
meaningful
descriptions
of
the
activities
which
demonstrate
how
the
group
project
was
progressed:
up
to
10
marks
• Peer
Assessment:
5
marks.
NOTE:
This
is
applied
to
your
individual
grade
and
will
be
reflected
in
grade
centre.