I am pursuing PhD in Computer Science. I am working on writing my PhD dissertation proposal.

My main contribution in my work is going to be about data classification and prediction model.

How should I phrase my potential contributions / tasks in which technical contribution is about clustering/classification and prediction?

I am also planning to propose potential performance improvement in clustering / classification techniques that currently exist in my domain. What all should I plan to achieve? How can I go about phrasing it in the proposal?

I know most answers cannot be specific to my question but any pointer towards direction that I should take...where to look for answer etc. regarding this is much appreciated...

  • Hi and welcome to Academia.SE! I've removed the questions about doing data analysis from your post, because these are off topic at Academia.SE. You may be able to ask them on Cross Validated instead. (Also, on StackExchange sites we ask that you write only one question per post.) – ff524 Jan 8 '15 at 23:18
  • 8
    any pointer towards direction that I should take — Picture me pointing at your advisor. – JeffE Jan 9 '15 at 4:58

"If the proposed research is successful, there will be three main contributions. First,.... Second, ... Third,..."

What you need to fill in is the specific nature of your contributions. It's not enough to say that your algorithms will be "better" or "interesting" or even "useful". You might use these questions to form a list of specific contributions:

  • Will you be discovering something new about the problem of data classification and prediction that no one knew before, or only understood poorly?
  • Will you be combining two or more methods that had never been combined before?
  • Will you be simplifying the problem in some non-trivial way? Or expanding it? Generalizing it? Making it more specific or concrete?
  • Will you be formalizing the analysis in ways that have not before? What will your formal analysis be proving?
  • Are you applying an algorithm or formalism from another domain, that has never been done before in this one?
  • Are you demonstrating through experiments or other empirical research the viability of theoretical methods that have little empirical support?

Finally, imagine that you have completed your dissertation, and maybe you have published articles based on it. Now imagine what other researchers might need to use your work and build on it, and therefore to cite it. Why would future researchers need to use your work or build on it?


one rhetorical framework you can use is this one: problem X exists; model Y has been proposed previously; i propose model Z, which solves X better than Y in certain cases.

JeffE is right to point you towards your advisor. helping you develop an answer to this question is a big part of her job.

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.