General Instructions

 Investigate the applied problem outlined in the project brief and write a short report (5-10 pages) on your analyses.

Note: It is the content that is important. Longer reports will not be penalized directly, but longer is not necessarily better (whereas concise is definitely better). See the Project Mark Scheme for further details.

 Your report should consist of an introduction stating the purpose of the analysis.

 This should be followed by a data and methods section describing the data and explaining why the method(s) you have chosen are appropriate and briefly, in your own words, how they function.

Note: You should give a more detailed description of any advanced methods (e.g. the M-level topic; methods beyond the scope of the module) employed and why they are appropriate (to demonstrate understanding).

 If the data is pre-processed in any way (e.g. scaling, normalisation) this should be explained and justified (at the appropriate place in the report).

 The analyses should be described in detail in a results section.

Recall: any tables, graphs or plots included should be carefully labelled and discussed.

 The report should have a concluding section in which you summarise and interpret your results. If appropriate, the results from different methods should be compared and any similarities or differences commented on. If appropriate, you should attempt to draw practical conclusions from your analysis.

 Tip: In general, you should not report the poor results of lots of classifiers that you have tried in the main text (see Mark Scheme). You may wish to include this sort of trial and improvement in an appendix (but if you do, it should still be presented in the formal report style).

 The report should have a future work (sub)section in which you suggest how the analysis could be extended and/ or improved.

 R code should be included in an Appendix.

(Note: This is included for two reasons: to check you can perform. this analysis correctly (so the marker may run your code to check it works as intended) and to identify errors if the results are not as expected (and hopefully still reward your efforts). Therefore, it is in your interest to include all R code clearly and concisely- see model solutions for good examples.)