A Template for Producing IT Research and Publication

1

• A Template for Producing IT Research and
  Publication
• Hsinchun Chen

2
Agenda

• Choosing Publishable Research
• Abstract
• Introduction
• Literature Review
• Research Questions
• System and Research Design
• Research Testbed
• System/Research Experiment
• Findings and Discussions
• Conclusions and Future Directions
• References
• Acknowledgement

3
Choosing Publishable Research

• A publishable paper needs
• an interesting research topic,
• good lit review,
• new algorithm/technique,
• relevant testbed,
• systematic evaluation (comparing with existing
  approaches and with statistical tests)
• Writing needs to be precise, concise,
  professional and entirely error-free

4
Choosing Publishable Research

• Research topic needs to be new and interesting.
• Avoid old and well-studied topic.
• Research could be technique/algorithm driven or
  application driven.
• Read a lot. Understand the current trends and
  directions. Need comprehensive lit review.
• Use well grounded methodologies.
• Compare with existing techniques/approaches with
  data sets.

5
Title

• 8 words or less.
• Develop a title after finishing the paper.
• Title needs to reflect the essence of the
  research.
• Dont use cute title, e.g., To aggregate or not
  to aggregate
• Use project/system acronym with clear relevant
  meaning, e.g., COPLINK, BioPortal not ALOHA.

6
Abstract

• Most important part of a paper the first
  impression!
• Abstract should reflect the entire paper.
• 200-300 words in one paragraph.
• 2-3 sentences to summarize problem motivation.
• 2-3 sentences to describe proposed method or
  algorithm.
• 2-3 sentences to summarize evaluation method.
• 3-4 sentences to summarize key findings.
• Write abstract after finishing the entire paper.
  Select key sentences from paper.

7
Introduction

• Motivate the research topic.
• 4-6 paragraphs. Less than 2 pages.
• Describe the importance of the research topic,
  current approaches, proposed methods, and
  structure of the paper.

8
Literature Review

• A major part of the paper. Need to show that you
  know the field.
• Read and digest many papers before writing.
• Need to review seminal works and critical new
  works (past 3-4 years in major journals and
  cnferences).
• 3-6 pages. 3-4 major subsections.
• Never do a laundry list review. Never do too much
  tutorial.

9
Literature Review

• Select only closely relevant works to review.
• Best to present as a taxonomy with 2-3 critical
  fields to classify works.
• User a detailed table to summarize and compare
  past works.
• Sometimes a good comprehensive review paper can
  stand on its own.
• Need to critique past works (critique not
  criticize!).
• Need to summarize research gaps that would lead
  into your proposed research questions.

10
Research Questions

• Summarize research gaps in the lit review
  section.
• Need 3-5 research questions that show the focus
  of research.
• Half a page.

11
Research Testbed

• Use research testbed to validate designs and
  approaches.
• What data sets will be used in the experiment or
  evaluation?
• Testbed should be interesting, relevant, and
  significant.
• For new emerging critical applications, research
  testbed could become the focus of research.
• 2-4 pages, in detail.

12
System Design

• Describe how the overall system/design works and
  its key components.
• Use an overall diagram with boxes and arrows.
• Explain the rationale of each component.
• 3-6 pages.
• Need to describe key technical innovation and
  technical details (algorithms, mathematical
  notations).

13
System Evaluation/Experiment

• A paper wont get accepted by a top-tier journal
  without a detailed, systematic evaluation.
• Present research hypotheses focused and
  measurable.
• Focus on the experiment and evaluation.
• Summarize evaluation methodologies adopted in
  past research.
• Compare with existing benchmark techniques.
• Describe experimental procedure, conditions,
  tasks, subjects.
• 3-4 pages, in detail.
• Need statistical tests (t-test, F-test).

14
Research Findings and Discussions

• Summarize key findings in a clear and
  understandable format.
• You may group your findings in chunks, each of
  which starts with a bold summarizing sentence.
• Present examples and discussions right after each
  finding to bring insight and better
  understanding.
• User screen shots to illustrate.
• 4-8 pages, in detail.

15
Research Findings and Discussions

• Tables and figures are critical.
• Need to be consistent and neat.
• Highlight interesting numbers.
• In caption, you may use 3-4 sentences to describe
  more details about a figure or a table.
• Use a small paragraph in text to explain the
  essence about a figure or a table.
• Need statistical tests (T-test, F-test). P value
  (significance level) at 5.
• Reject or accept hypotheses postulated.

16
Conclusions and Future Directions

• Can have some duplication with the abstract.
• Summarize key findings.
• State the contribution, but dont overstate it.
• State research caveats.
• Dont mention trivial future directions.
• Point to several promising directions.

17
References

• Where has similar work been published?
• What kind of articles are accepted by the target
  journal? Reference related papers that were
  published in the target journal.
• Must have 5-10 key journals, key conferences, and
  key authors in the field.
• Number the references in a consistent format.
• Include a few of your own works. Avoid having too
  many self-citations.
• 15-50 citations for most journals.

18
Acknowledgement

• Funding sources and grant number.
• Research partners and participants.

19
Additional Notes

• Conference papers only need to have about 50 of
  the above coverage.
• Target only major professional conferences to
  report work in progress and solicit feedback.
• Conference papers need to be quickly expanded to
  target at top-tier journals.

20
Additional Notes

• Select the right journals for publications.
• Research journal scope, past publications, and AE
  background.
• Consult senior colleagues for suggestion.
• Behavioral-oriented MISQ, ISR
• Math-oriented MS
• System-oriented JASIST, DSS, JMIS
• Algorithm-oriented IEEE TKDE, ACM TOIS, IEEE SMC
• Application-oriented CACM, IEEE Computer