Anatomy of a Research Paper (Part I)

This is a simple structure I follow for any conference paper.

Abstract

• 1-2 sentences: Why are you doing this?
• 1-2 sentences: What are you proposing, and what it does.
• 1-2 sentences: What experiment you did and what are the results
• 1-2 sentences: What is the finding after comparison against existing approaches?
• Always write the abstract after completing the whole paper.

Introduction

• First paragraph====
• What problem are you trying to solve?
• Who face this problem? Is it really an important problem to solve?
• How to solve the problem? Is your proposal is a potential way to solve the problem?
• Second paragraph====
• What are the existing approaches that tries to solve the same problem, and how they try to do that (using short hints)?
• If there exists no such study, then your approach is a novel one. But think twice, does your approach really make sense? Possibly talk to another person and get feedback.
• If there exists some studies, what are their limitations and restrictions?
• Use inline numbering to point out the limitations
• Third paragraph=======
• What system/approach you are proposing?
• What it does, and solves the mentioned problem in para 1?
• Does it solve the limitations of the existing approaches? If yes, you are on the right way.
• Use inline numbering to point out how it solves the limitations.
• Fourth paragraph=========
• Experiments: what data set you used, and what are the results?
• Did you compare against the existing approaches? What are the findings there?
• Are the overall results are promising? If not you can consider to repeat the experiment to reach a satisfactory level.
• You can consult the literature to check whether your results make sense or not.
• Last paragraph====
• Mention rest of the sections in the paper.

Motivating Example

• Show the most appealing feature of your proposal, may be with a diagram or snapshot of your tool.
• Start a problem scenario with the user like a story.
• Show how the problem cannot be solved by existing approach.
• Show the problem can be easily solved by your approach.

Background

• Think about the theoretical concepts your had to learn for this research.
• Also add other theoretical topics that you know clearly, but it will help the reader to understand your work.
• Most of the concepts will be found in Wikipedia. You can learn from the site, but use your own language to describe the concept.

Proposed Approach

• This section will contain different parts of your proposed solution.
• Let me guess, you are proposing an algorithm (theoretical) or an approach (not purely theoretical) or a prototype that implements the approach or a pure tool such a visualization tool.
• If you are proposing an algorithm develop a flow chart, if you are proposing an approach make schematic diagram.
• Then discuss your approach or algorithm step by step in subsections and refer to the diagram.
• If you are proposing metrics, they should be before the main algorithm.
•  Also discuss how the result of your algorithm will be presented to the users.
• If you are proposing a tool, make a block diagram or a snapshot that shows different working modules of the tool.
• Discuss different features of your approach referring to the module in the diagram, and solve different example problem scenarios.
• Here the scenarios should be more technical, detailed and entertaining.
• If are proposing an approach with a prototype, then provide a short overview of the tool in the first subsection, and then focus more the approach. The tool does not exist without the approach.

So far you have everything that your are proposing. Now comes the experiment which will show whether your IDEA works or not.