As the title states I'm curious how to justify the choice of baseline methods for academic paper when proposing a novel algorithm/method. For example, when proposing an algorithm, there is a suite of methods applicable. When I choose one method to compare my algorithm to, how do I justify it?

  • 3
    Welcome to Academia SE. Can you please specify what you mean by baseline method? Also, if I am guessing correctly what you mean by that, I find the question rather difficult to answer except in a maximally fuzzy way, as it always depends strongly on the actual reason for your choices.
    – Wrzlprmft
    Sep 28 '15 at 21:31
  • 1
    The justification of the methods you use should start well before the paper - what did you do, and why did you choose to do that rather than something else?
    – Jon Custer
    Sep 28 '15 at 22:37
  • I believe that OP could clarify, but they have a new method that they want to compare against a baseline and are curious about the best way to find which baseline to compare against. In my field(s) this would be the fastest, most accurate method already published.
    – Bill Barth
    Sep 28 '15 at 22:50
  • 1
    You may lift the hold. The question seems pretty clear to me. Eg., if there are a list of classifiers in machine learning (RBF, ANN, GA, SVM, etc.), and you choose one, how to justify your choice of algorithm?
    – Ébe Isaac
    Sep 29 '15 at 5:24

When you propose a method in a research paper, it is customary to describe how you came up with the method and why choose it over the others available in literature. There are a few tips you could follow to justify your selection.

  1. The traditional way to do this is by doing a thorough literature survey of all other methods assessing the advantages and shortcoming of each and then show how your chosen method could probably supersede those methods.
  2. Sometimes you may find many such methods that are applicable but for which you cannot find existing results. In such a case, the best thing to do is to produce results for all of those methods and select the one with least drawbacks and most desirable merits. If the amount of publishable work exceeds your paper limit (or if your time is limited), choose one among such methods for your current publication, but discuss the other methods that can also be applied that could be of scope for future work for research.
  3. A few journal publishers would also welcome negative results (such as PLoS ONE). Hence you could assess a method and show how suitable/unsuitable it is for a certain application based on the results.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.