Guidelines for writing the final report

The report should be written in the form of a paper.

The following sections should be included:

The report should have between 10 and 30 pages written in Times New Roman (or similar), 12 point size, single or 1 and 1/2 spaced. All margins should be less or equal to 1 inch.

This is what the sections above should include:

The abstract should be  one paragraph of up to 300 words. The abstract should summarize the content of the work presented and underline its importance. The reader should be able to assess the importance of hte work and the original contribution from the abstract. There are publications that only include abstracts of papers. You put here whatever you  think would entice the reader to read the rest of the paper as well. The introduction should present the context, the problem and its significance. Not more than 1 page. A review of  what other people have done related to the problem studied.  You should go into some detail for that particular part on which you are building. Present advantages and disadvantages or all existing work. Use complete references. The purpose of this part is to  convince your reader that: i) you know what has been done in the field about the problem studied and ii) there was a need for the work you are presenting. Not less than 2 pages. Use as much  space as needed. This is the section in which you present your work. Make sure that you do not assume anything about the reader's background. Make sure you present all that is needed for somebody with a   general scientific background to understand your work.
Do not use statements like "it is obvious that...", "it is clear that...", etc. Support each individual statement with arguments and/or references to existing results.
The technique that you are proposing tends to be always the best there is. Try to be objective. Discuss advantages AND disadvantages. Justify why you have  designed the experiments as you did. Present the hypothesis to be tested by the experiments and anticipate what would prove or disprove them.
Present the experiements in details. Make sure that all the information needed to reproduce your results is there. Present you conclusions. Describe again but very briefly what the problem was, how you addressed it and what the main results are. A reader should be able to  get a good idea about the paper just from the  conclusions section. The references must include all the information necessary to retrive that piece of writing from wherever it is. Web pages and "personal communications" are not acceptable as references.

These are examples of references for a journal article, book and conference paper, respectively:

Draghici S. - A neural network based artificial vision system for licence plate recognition, International Journal of Neural Systems, pp. 113-126, vol. 8, no. 1, 1997.

Preparata F., Shamos M. I.  - Computational Geometry - An Introduction, Springer-Verlag, 1985

Draghici S. - On the Complexity of VLSI-Friendly Neural Networks for Classification Problems, in R. Mercer, E. Neufeld (Eds.), Lecture Notes in Artificial Intelligence 1418, Proc. of AI'98, The XII Canadian Conference on Artificial Intelligence, Vancouver, British Columbia, Canada, 18 20 June 1998, pp. 285-297, Springer-Verlag, 1998.

This is how a reference should appear in the text [Draghici, 1998].

If you are writing a review paper, the structure is a bit different.
 

You give a general image about the field presenting the things that are important across various techniques discussed. You also discuss the  criteria that you are going to use to organize the rest of the paper. These  criteria will induce the remaining headings.
In a review paper you have to:
identify some criteria that allow you to put some order in the chaos of the techniques reviewed
organize the various existing techniques using the criteria above
describe the main technique and ideas so that one could actually understand them.
give at least one reference for each technique or idea mentioned.