I build an predictive machine learning model that generate the probability to default over the next 2 years for all the companies in a specific country. I used for training the algorithms financial data for all these companies and also the NACE codes (domains of activity) and I'm wondering if I will develop a better model if I somehow segment the population of B2B in segments and run distinct models on these segments.
Hope you can advice!
Lots of thanks in advance!
Download the attachment book from the z-library. Best wishes David Booth
It depends on how powerful you want to make your model. Intuitively one may expect better granularity by segmenting ,this could be an overlay on the data ,but a general clustering with a post segmentation might be more appealing. As an example face recognition is more appealing as an application if it is general , rather than segmented countrywise or racially.
Good Luck
Cheers
Dear Coralia Zotic
Demographic characteristics are the most essential in B2B market segmentation, followed by operational variables. These capture the most important qualities of potential target clients and describe their significant needs.
Age, gender, family status, employment, income, race, and religion are all factors. Marketing to demographics allows you to connect with your clients more effectively. Geographics: The primary variables that effect your clients' demands with this form of segmentation are region, climate, and population density.
Kind Regards
Qamar Ul Islam
I just came across a very interesting maths development which might help in navigating your problem. As I understand ,from this reference below , it should be possible to break up your data ( segment) into meaningful parts . This may require formulating the problem into a structured set of vectors.
https://www.quantamagazine.org/computer-scientists-solve-kadison-singer-problem-20151124/
Cheers
You can work on different aspects starting as cited A demographic approach and it could be also a geographic one depending on what information the dataset includes also you can try to identify behavioral patterns within your data. Also, you could go further by focusing on Customer capabilities & the Nature of the existing relationships.
- Badges
- Science topic