Before you ask what sections should be contained in a survey paper, you should first understand what is a survey paper. What it is *not* is simply a core dump of a bunch of papers in a common area.

Think of a survey as a research paper whose data and results are taken from other papers. This means that you should have a point to make or some new conclusion to draw. And you'll do so by collecting data from a broad collection of previous works.

As a previous responder has said, you should have a thorough and deep knowledge of the field that you are surveying. This knowledge should be sufficient to be completely aware of the the main themes, directions, controversies, and results in that field.

The point you will make will determine the organization of survey paper. The structure of the main sections of the paper will reflect the structure of field. Some possible example structures (which of course depend completely on the topic) might be:

1. Increasingly complexity or scale: There may be a spectrum of solutions and you might organize them by complexity or scale.

2. Static vs. dynamic: Many field organize by static techniques, dynamic techniques, and even hybrid.

3. Partitioning the design space: Lots of systems are made up of components, so maybe for an compiler paper, you could divide by the classic scanner, parser, symbol table, code generator and optimizer.

4. Major techniques in a field: For example, in fault tolerance, you see fail stop vs. fail forward, or logging vs. hot-backup. In concurrency control, there is a natural divide between optimistic vs. pessimistic techniques.

5. Historical: sometimes the course of development of a field has a clear linear nature and is intrinsically interesting in itself. This is an over-used techniques in many cases where it really doesn't add understanding.

There are lots of possibilities for a given topic and it is this organization that is the hardest part of writing a survey paper. (I'm sure that many of you can give good examples of organizations that have worked well for you.)

You'll have written a successful survey paper if you can communicate not just the list of results, but more important, your understanding of the structure of the field.

This is a high bar to set. And it is also why I never ask students in my graduate classes to write such papers; they just don't have the experience and perspective to write a good survey.

Collect as many papers as you can. Read them and note down their features and limitations your own words and sentences. Follow IMD format (Introduction, Methodologies, Discussion).

No need to add your own your, but in case if you have done enough work in that area you can very well discuss over it.. other than study papers as much as you can in the field that you want to write a review.. and write a clear report That will do.. God bless U...

You need to mention a inclusion/exclusion criterion for the selected papers

Rafael Capilla, Sir please explore more what do you mean by inclusion/exclusion criterion

Giftson Sam, Sir I have sorted out the work done in my research area now how to give it a form of review paper. I have not done enough work just only searched the work done till date.

So, I mean that if you will review dozens of papers, you need a criterion to decide which papers will be included in your survey and which ones not. You need to say if at a first glance you will select papers based on, let's say, title+abstract, and after in a 2nd round you refine your selection by reading the complete paper or using a different criterion. In the end, you will compare in your survey the selected papers. You can download any survey to take a look to the structure. Cheers

It's a big work. There're already many reviews in a lot of fields. Gather the good ones (if exhaustive), do a survey from them, and read/evaluate the new papers since the last good reviews.

My experience is that writing a survey paper is more difficult than writing a research paper. One needs to have many rounds for modifying it. The first thing is that one should have a good collection of papers. I suggest that one should concentrate on papers published in top journals and top conferences (preferably tier 1 for both). Try to understand them at least the major idea, focus, features and limitations. One should be broad minded i.e. one should not be impressed by a research paper nor should one dislike a particular paper. Consider yourself to be an expert in the field but write politely and in normalized fashion. One should be careful that he/she is reviewing the the work presented in papers and not the authors.

If you have a well defined research question, you can follow the Kitchenham guidelines to do a Sistematic Review (writing a protocol including the search string, inclusion/exclusion criterias, etc.). If you have many research questions or you want to explore a broad field, then you can do a Mapping Study, which is a kind of bibliography research with a more light protocol.

In both cases, Kitchenham and others have guidelines in several papers so I think that you will not have any trouble doing it. Ask for help if not.

Dear Sudhakar ,

I think, this is very intillegent quations,

First: You can put collection of Quations? What is the problem u need apply survey of it.

What are the main parmaters or variables of that problem? Is the problem solved or remined as open still problem yet? Can i using that survy in the future as guide to bulding new ( tools or algorithm ) or to developing (priviouse methods or theoritical Fundementals).

Second: search the web and book about the papers has the same problem defination in point number first and order that papers base on date (from last to up date).

Third: for each paper, determind the following points(methodology used in it to solve the problem, theory, preprocessing type, The problems that the paper can be solve The problems that the paper can not be solve"remind problem", and Accuracy of the results)

Fourth: Compair among all the papers based on the points determined in the above point.

Finally: Disuess this results that expliend in the table that you build and show your' opnin about this results and Give the reader summary about the( simlarity and desimliarty rate) among the papers used in that survy.

Note: more new ask can be add to third point.

Best Regards,

Samaher.

In the survey paper, you must to have abroad knowledge on the topic and then present the materials in an analytic view. In the meantime you have to highlight the main challenges and trends.

Some details taht can be useful to the readers of an survey paper:

1- Cite together papers dealing with the same method of discovering association rules. Some present the first draft of the methods, another the mature method, others new applications of the same methods. It is easy, generally those papers share at least one author.

2- Provide one or more taxonomies of methods for association rule mining. Each taxonomy should be based in some important criteria (for example, classification of methods by algorithm, or by knowledge representation, or by problem niche addressed). May be there is a taxonomy already created, then you should group papers into classes; may be there is no taxonomy, and you should propose one and then you shoul group papers into classes. It is easier if you did step 1.

3- Discuss advantages and shortcomings of each method.

I suggest to review ACM Computing Surveys, there you can find many survey papers: http://dl.acm.org/citation.cfm?id=J204&picked=prox

I recommend you to follow the example of this fine survey paper: A survey of online failure prediction methods, http://informatik.hu-berlin.de/Members/salfner/publications/salfner10survey.pdf

From the ACM Computing Surveys Editorial Charter (http://csur.acm.org/charter.html):

"Basically, a Computing Surveys article answers the questions, 'What is currently known about this area, and what does it mean to researchers and practitioners?' It should supply the basic knowledge to enable new researchers to enter the area, current researchers to continue developments, and practitioners to apply the results."

refer the existing research paper related to the topic. Conclude each paper and rewrite with your topic

One have to find, sort, arrange or integrate the works done since inception in particular research area and present his/her own perception with conclusions. It is not necessary to add his/her own experimental contributions. I want to confirm that am I following right ?

This is an excellent question. It will be useful for many researchers. First of all, one has to be clear about whatever one wants to surrvey or review about. Write on something that one is the expert on thé field and has some experience IN doing research in. Do not Forget to included one's research work or finding in the Discussion Section.

In the RJAS journal (link: www.rsu.ac.th/rjas) one review article in various field is included at the end Of every issue.

N. Jan, M. Ibrar, Systematic Mapping of Value-based Software Engineering - A Systematic Review of Value-based Requirements Engineering. You can find all steps reported in this study. You can find all important steps ................. I hope You will like it.

Sir, for review paper, you should sufficient knowledge in your field along with publication. This types of paper are same as research articles with little difference. You may visit any reputed journal for review publication.

Thanks

Dear Sudhaker, first of all, find problem statement in cloud computing, In cloud computing many areas are there, exactly what area you are looking for, then, search all the reputed journal paper and collect atleast 50 papers till today,read thoroughly and make summary, thereafter,what dataset used and techniques applied thereof, again make a summary, Atleast your initial problem will be solved.You can follow my review paper related to software engineering problems:The application of intelligent and soft-computing techniques to software engineering problems:a review. I think, it will be sufficient input to you.

Thanks ganesh deka for nice content answer and chatchai for good question

Thanks Ganesh Deka. Your suggestions have helped me resolved the problem of how to design an Evaluation Form for A Review Article sent for publication in the RJAS journal at the following link: www.rsu.ac.th/rjas . Ar present, we use tehIntroduction, Materials and Methods, Results, and Discussion style which many Reviewers have complained on its irrelevancy.

Before you ask what sections should be contained in a survey paper, you should first understand what is a survey paper. What it is *not* is simply a core dump of a bunch of papers in a common area.

Think of a survey as a research paper whose data and results are taken from other papers. This means that you should have a point to make or some new conclusion to draw. And you'll do so by collecting data from a broad collection of previous works.

As a previous responder has said, you should have a thorough and deep knowledge of the field that you are surveying. This knowledge should be sufficient to be completely aware of the the main themes, directions, controversies, and results in that field.

The point you will make will determine the organization of survey paper. The structure of the main sections of the paper will reflect the structure of field. Some possible example structures (which of course depend completely on the topic) might be:

1. Increasingly complexity or scale: There may be a spectrum of solutions and you might organize them by complexity or scale.

2. Static vs. dynamic: Many field organize by static techniques, dynamic techniques, and even hybrid.

3. Partitioning the design space: Lots of systems are made up of components, so maybe for an compiler paper, you could divide by the classic scanner, parser, symbol table, code generator and optimizer.

4. Major techniques in a field: For example, in fault tolerance, you see fail stop vs. fail forward, or logging vs. hot-backup. In concurrency control, there is a natural divide between optimistic vs. pessimistic techniques.

5. Historical: sometimes the course of development of a field has a clear linear nature and is intrinsically interesting in itself. This is an over-used techniques in many cases where it really doesn't add understanding.

There are lots of possibilities for a given topic and it is this organization that is the hardest part of writing a survey paper. (I'm sure that many of you can give good examples of organizations that have worked well for you.)

You'll have written a successful survey paper if you can communicate not just the list of results, but more important, your understanding of the structure of the field.

This is a high bar to set. And it is also why I never ask students in my graduate classes to write such papers; they just don't have the experience and perspective to write a good survey.

1- http://www.ijarcsse.com/docs/papers/january2012/V2I1074.pdf

2-http://www.ccis2k.org/iajit/PDF/vol.11,no.4/5426.pdf

3-https://www.facebook.com/pages/Data-mining-and-cloud-computing/224219914263598

Enjoy

@Barton Miller, thanks a lot sir

In reference to writing a review article, I prefer Ganesh Deka's comments the most. It is concise and conclusive in itself. And if you can write a Discussion Section by following the following Discussion rubrics (see the attachment) developed by Associate Professor Dr. Nikhom Tadaeng, Bangkok, Thailand, the article will be worth to read. Those given by Barton Miller are good for a survey article though.

whats about this this image ? Can you please explore it more and add the you tube link of this video ?

The Youtube link VDO was given in Thai language. The developer, Associate Professor Dr. Nikhom Tadaeng, called it the HONE x RISCA Rubrics. The full term for each abbreviated letter was annotated. For example, H was a shortened form of Hypothesis; and R from Rationale/Relevancy. So, one should sequencing the order of Discussion Topic from the H X R square first, and then either by H x I or O x R, etc., where most appropriate or relevant. Usually, one can discuss on only around topics for 5-10 squares the most.

May you paste the link here please

Dear Sudhakar Singh, Sir

I have added the requested link for you below. In the VDO, Dr. Nikhom Tadaeng gave the advice on how to write a research paper Discussion Section. But I also recommended using this technique to write a review paper as well for authors who submit the manuscript for publication in the Rangsit Journal of Arts and Sciences (RJAS) at URL/Link: www.rsu.ac.th/rjas

I have talked to Associate Professor Dr. Nikhom Tadaeng and asked him to give permission to me to adopt and promote his technique about a year ago. He kindly gave the permission for me to do so. I felt very touched and much obliged to his contribution.

It would be a great honor to us if you would consider Login, read, download free full papers & promote our journal at the RJAS Website at www.rsu.ac.th/rjas . the file HOW TO LOGIN RJAS Website was also attached to guide you how to login.

http://www.youtube.com/watch?v=ALdJJyD93IY

It would be a great honor to us if you would consider Login, read, download free full papers & promote our journal at the RJAS Website at www.rsu.ac.th/rjas . the file HOW TO LOGIN RJAS Website was also attached to guide you how to login.

Thanks Ganesh Deka. Your suggestions have helped me resolved the problem of how to design an Evaluation Form for A Review Article sent for publication in the RJAS journal at the following link: www.rsu.ac.th/rjas . Ar present, we use tehIntroduction, Materials and Methods, Results, and Discussion style which many Reviewers have complained on its irrelevancy.

I think that a good a survey paper and review paper at least include the following important sections:

1: Introduction: the background information, importance of the study

2. Materials and Methods: summarizing the materials and methods used

3. Results: summarizing the main results achieved

4. Discussions: Analysing and discussing the results in details

5. Conclusion: Brief summary of the results, comments, recommendations, outlook. etc.

@ Shouke Wei, Please let me know that in review paper how can one write results without running any experiment ???

Since it is a review paper, one just summaries the results in the papers that are reviewed. If you do any experiment, then it is a research paper.

@Shouke_Wei Yes, the author of a research paper only needs to summarize the results of other people's papers... if the results are explained in detail. Sadly, many papers present results in a very abbreviated form (example: Proceedings papers with a thigh page limit). Without detailed explanation (ex.: parameters used, configuration of the machines running the experiments, quality of the data, etc.), comparison between results of 2 different papers is quite fuzzy. Sadly many proceedings does not consider this fact. So, a good advice would be to get in touch with some authors and ask them for more information about their results.

While It is always a good idea to ask your PhD/Master students to write a surevy paper as a start point, it is also difficult for new research students to produce high quality surevy papers. In my experiences, a typical surevy paper should consist of the following, but not limited to, key sections at least:

1. Intruduction: to give background and motiviation why to write this surevy paper.

2. Scope and Methodology: to provide a framework/taxonomy to define the scope and methodology, techncail notations that you are going to use to conduct the surevy - A digram and/or a table for technologies catelog/classifiction would help a lot here.

3. Main body: to list, describe and compare the leading work in the areas using the uniform surevy method/style that your have defined in the above section. The numbers of sections and subsections are subject to your research areas and where to submit. The bottom-line is you should cover the top work conducted by the leading researchers in the area;

4. In end of each section and end of this paper, it is always a good ideas to summarize your surevy by listing the technologies/methods that you have discussed and compare them using a table or figure.

5. Again, it is always valuable for a good surevy paper to point out not only advantage of these studied technologies/methods, but also their limition so that you can provide/predicat some possible research directions for the students themselves and readers.

Here are 2 survey papers related to my areas for your references:

- Jinhui Yao, Shiping Chen, David Levy: Accountability for Service Compliance: A Survey. IJSSOE 3(1): 16-43 (2012)

-Jordi Guitart, Jordi Torres, Eduard Ayguadé: A survey on performance management for internet applications. Concurrency and Computation: Practice and Experience 22(1): 68-106 (2010)

Term Paper General Guidelines:

 Each student should conduct an in‐depth study of an advanced software Project Management

topic of his selection. The findings should be reported in a form of a publishable work.

 Use recent IEEE & ACM publications as references.

Submission Guidelines:

There are six main phases for this term paper, these are:

 Phase 1

 Proposal: general problem statement and initial set of references (after obtaining the

instructor’s approval).

 Phase 2

 Literature Review: Summary of the approaches presented in the references investigated.

 Phase 3

 Evaluation criteria/attributes: you are expected to identify a set of evaluation attributes

that can be used to compare the different approaches surveyed.

Or

 Research approach & data description: you are expected to identify the research approach

and the data to be used in your study.

 Phase 4

 Comparison: using the identified attributes, compare the different approaches surveyed.

Or

 Empirical study: Using the identified approach and data conduct your study.

 Phase 5

 Final Report: the final form of the term paper should include:

 Literature review, the identified evaluation attributes, the analysis of the

approaches surveyed, and some specific open problems that none of the

approaches surveyed have solved.

Or

 Literature review, the identified research approach and data, the analysis of the

empirical study, results and conclusion, and some specific open problems that can

be further investigated.

 Phase 6

 Presentation: You are expected to present and discuss your findings.

Thank you all, Got very nice insight to, how to write good survey paper .

1. Try to find similar studies on particular area selected for survey/review

2. select relevant papers from reputed repositories: IEEE explore, ACM, Science Direct, and others.

3. Try to use highly cited papers

4. use thematic approach by selecting appropriate categories and cluster the papers in such themes.

5. design a review methodology along with classification diagram.

6. write it as much as u can collect, and finally filter the required content .

7. compare the results of similar studies and arrange them in logical way in tabular form

There are some very well cited (around 2k citations, depending on who and how you ask) articles on how to conduct a good literature review. I found these two inspiring and helpful.

Webster, J., & Watson, R. T. (2002). Analyzing the past to prepare for the future: Writing a literature review. Mis Quarterly, 26(2), XIII–XXIII.

Kitchenham, B., & Charters, S. (2007). Guidelines for performing systematic literature reviews in software engineering. EBSE Technical report.

Thank you to all contributors on the topic, I'm collecting a number of papers for my survey and this discussion has become a great help, cheers!!! 

Thanks

I Thank all d people  fr giving their valuable suggestions in preparing a survey paper....

Kindly share the template for the survey paper

thanks for all, the discussion was also helpful for me.

Thanks all. Very good advises.

Please follow the suggestions in previous answers. They ARE good.

Additionally, a major purpose of a survey paper is classification of existing studies e.g., to prepare a taxonomy. Also, in case there is a debate on an issue, review the arguments of both sides succinctly. 

not easy

Survey paper or review article writing has never been an easy task. Only knowing the headings in the survey paper is not enough. You are actually required to fill in those headings with useful, logically arranged and clear material. I have been writing surveys from 9 years and have published 7 of them in top journals. I must share a source with you people which I found most useful for this purpose. http://researchpedia.info/steps-to-write-a-survey-paper-review-article/

Its the best ever steps or tips with detailed guidelines for each section. I hope It will help you all and your students or colleagues.

This type of paper just provide a review or reports:

1. Select a topic and collect multiple papers as many as you can. (use Google Scholar, ACM and IEEE)

2. Read Abstract, Related work and Conclusion first and follow step 3.

3. Now make a literature review: i preferred to use a excel sheet with columns ( Paper Name, Problem, Solution, Findings, Future work, Related work).

4. Then start writing paper.

5. Use tabular form to provide a comparative study of all papers.

https://www.cse.iitb.ac.in/~uday/soft-copies/writing-a-good-paper.pdf

This thread was helpful. Thanks for the discussion.

Clear and helpful, thank you sincerely.

follow advised of Shiping Chen

This thread was helpful. Thanks for the discussion.

These responses were really helpful, thanks for the discussion.

A survey/review paper is not just a bunch of random references about a certain topic. You should have a deep knowledge of every aspect regarding the subject. Then, divide it into different parts/areas, and try to outline the ones in which the readers might be interested in (and which are not widely covered by previous literature). These will be the core of the manuscript from introduction to conclusions. One or more case studies, presented along the manuscript, might also help for justifying the reasons for focusing on particular aspects.

That was helpful, thank you.