A short guideline specifying the possible structure of your paper

1. Title:
Select an appropriate expression describing precisely your research domain and your main objective / conclusion

2. Abstract:
States the problem, the idea, your claims and some results

3. Introduction:
States the problem
Why it is important to address it
What is presented and why
What is not covered / limitations of your approach and why

4. State of the Art:
Indicates the most important previous work
Summarizes them briefly
Only relevant and related work
Useful to see the limit of current practices
When well done, the new suggested method will introduce itself effortless

5. Suggested Approach / Formalism:
Defines clearly the notation used
Introduces only the needed elements
And one at the time (step by step), top-down
Uses an example (not too simple, not too complex) to help the reader (running example)

6. Methodology / Evaluation:
Presents some evaluation / comparison with previous work or against a baseline
Shows how general the new suggested formalism is (or whether current formalisms are special cases of the new model)
Used available benchmarks (or sometimes simulation can provide an answer)
Statistical interpretation where appropriate

7. Conclusion:
Recaps the main idea of the suggested approach
Main results found
Open new directions (improvements / applications, etc.)

More information:
– Zobel, J. Writing for Computer Science. 3rd Ed., Springer.
Available online