Student’s Name Module Code COMP312P
General Coments:
1. Methodology Mark (%)
The methodology is as simple as could be or is mising
entirely. The wrong type of algorithm ay have ben
used and demonstrates a lack of understanding of the
problem.
The methodology is novel and draws in extra
information in an intelligent way. The description of
the methodology displays a ken understanding of
the problem.
2. Evaluation Strategy
The evaluation strategy is either mising or
inapropriate. Erors in the strategy have gone un-
noticed leading to a biased model.
Excelent evaluation strategy that covers all the bases
and is well justified.
3. Presentation of Results
No atempt has ben made to present and discus the
results. They are difficult to understand and no
attempt has been made to interpret them.
The results are wel interpreted revealing insights into
the data and the aproach. Excelent use of statistics,
tables, and graphs.
4. Interest of Aproach
No interest or novelty. No extra sources of
information is used and the aproach is simply a like
for like copy of an existing solution.
A novel approach that demonstrates an excellent
grasp of the problem. Extra sources of information
are used well.
5. Format, structure, references, and clarity of writing/code
The submision was in the wrong format or large
parts were mising. No references and the writing
and code quality are por.
Superbly laid out notebok with clear logical steps.
The notebok runs and the code and surounding text
are well writen and easy to understand. The
references show an impresive level of background
reading.
Overal Mark
COMP312P Asignment 2
Assesment Form. (2016)
Marking Descriptors (Underline coresponding descriptor)
Range Descriptor
90-100%
Exceptional
• Methodology is of publishable quality. Developed new approaches to tackle the problem
• Evaluation is constructed in such a way as to remove any positive bias. Parameters have ben corectly set and an
appropriate re-sampling method has ben used and justified. Goes well beyond what was expected
• Results are clearly presented and go into exceptional detail. There is briliant use of tables and graphs and the
surounding text is excellently written and shows real insight into both the data and the approach
• The approach is completely novel and draws in extra data in an interesting and novel way
• There are no flaws in the speling, the references show a god range of background reading, and the writing and
code is of publishable quality
80-89%
Outstanding
• A methodology that shows novelty and is clearly described. Multiple data sources are used and the model brings
together multiple inovative strands
• The evaluation strategy compliments the methodology and is logically explained and justified. Al relevant
parameters have been estimated
• Results are excelently displayed with apropriate visualisations and reporting used. The surounding text shows an
excellent level of critical understanding
• There is novelty and inovation in the approach. Impresive use of outside information
• Very few flaws in the writing, layout, references (of which there are a god amount), and code
70-79%
Excelent
• An impresive methodology that uses multiple sources of information. The reason for using them is wel justified
and advanced methods have been used
• Excelent evaluation strategy with al of the basics wel covered
• Results are clearly presented and some atempt at interpreting the results have been made
• The approach is interesting and the reasons justifying the aproach are wel made
• Very god write-up with very few mistakes. Code is generaly wel commented and there are some references
displaying background reading
60-69%
Good
• A nice methodology is presented and an extra data source has been used. The use of extra information is
appropriate and well justified
• The evaluation strategy is logical and wel reported but might contain some errors
• Results are al present but perhaps not formated in the optimal way (i.e. tables and graphs not properly used)
• The use of extra data or modification of the core model has displayed novelty and interest
• Well writen with good code and structure. Some references.
50-59%
Satisfactory
• The methodology is appropriate but porly described. No extra data sources have ben used
• An evaluation strategy has been employed but there are flaws in the estimation of parameters
• The results are generally wel presented but some key information is mising
• There is some novelty to the approach but it is litle more than training a standard classifier
• Write-up is generaly OK but there are a few speling errors. Some references, adequate code and coments
40-49%
Passable
• A methodology is presented but it does nothing more than train on the raw features in the training set
• The methodology is evaluated but no appropriate re-sampling methods are used or there are errors in the aproach
• Results are poorly presented with key information mising
• There is no novelty to the approach
• Some flaws in the write-up with parts mising, no references, or poorly writen documentation and code
20-39%
Unsatisfactory
(Clear fail)
• An atempt at a methodology has been made but it does not work
• No real atempt at an evaluation strategy or there are major flaws in the aproach
• Results have not been properly presented or there are major erors/omisions
• No novelty to the aproach
• Poorly writen documentation and code. No references
0-20%
Hopeles
(Unaceptable
fail)
• Methodology is mising
• Evaluation strategy is mising
• Results are mising
• The approach has been copied from another source without reference
• Seriously flawed write-up with litle to no documentation, litle to no code, or submited in the wrong format