1

I am writing my masters thesis at a company. The thesis is about customer segmentation. One research contribution of the thesis was to propose a way to validate the found clusters in an unsupervised setting.

That being said, to find clusters, I followed a very standard process. From raw data, I made a feature table, did some data pre-processing, dimension reduction and then clustering. I did not have to do any literature review for that.

Also for the cluster validation, me and my supervisor kind of came up with the method.

Now that I am writing my thesis, I don't have anything to write in this 'Related Work' chapter. Or it seems to me like that way. There has been some customer segmentation related work done at the company before, but there is not any scientific work. Can I mention those works?

2
  • 3
    Did you do a literature review? That would be the place to look for "related work."
    – Buffy
    Aug 22, 2022 at 13:48
  • 1
    "Related work" is usually meant a bit more broadly- surely there are some academic or otherwise technical publications that talk about the state of the art for customer segmentation. I give simple versions of this as examples in some of my undergrad courses.
    – user137975
    Aug 22, 2022 at 23:00

2 Answers 2

4

From the question I see at least three opportunities where including a discussion on related work would be useful.

One research contribution of the thesis was to propose a way to validate the found clusters in an unsupervised setting.

In order to qualify the proposed method as a research contribution, it would be very useful to include other existing methods on cluster validation and describe why they do not meet your needs.

From raw data, I made a feature table, did some data pre-processing, dimension reduction and then clustering.

While you might consider such process as standard and thus not requiring any literature review, it will be useful to include a description of the "standard" machine learning lifecycle* and the theory behind the methods applied in each step. This will help demonstrate you have mastery of the materials.

There has been some customer segmentation related work done at the company before, but there is not any scientific work.

While previous approaches may not be scientific, they are nonetheless approaches. Presumably, the project you are doing is providing some extra value to your stakeholders and thus it will be a useful discussion on how is it doing so.


* Spoiler alert: A "standard" machine learning lifecycle is less standardised than you think.

2
  • Lots of thanks for your answer. It helped me when writing :) Sep 7, 2022 at 10:27
  • I do not think it was good, but it will do. I think. Sep 7, 2022 at 17:55
0

It is common to feel like you have nothing to say in the "Related Work" section of your thesis. After all, you are writing about your own original research. However, it is still important to situate your work in the broader context of existing knowledge. In this case, you can mention the customer segmentation work that has been done at your company. While this work may not be scientific, it will still be helpful for readers to see how your work fits into the larger landscape. Additionally, you can use this section to discuss any related work that you have found in your literature review. By situating your work in the broader context, you can help readers to understand its significance and contribution.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .