Model Machine Learning RRS feed

  • Question

  • Jaime is greeted by a student who just started with the predictive models of machine learning.

    I wish you could give me some guidance on some of the questions I have about predictive models using Azure Machine Learning.

    I commented that I am doing my thesis project and I want to apply azure machine learning, basically it is a project applied to the control of medical visits made by the salespeople and what I want to propose is that the predictive model can return a program of weekly medical visits where different variables are handled such as the place of visit, contact or doctor to visit, the time of visit, the products to promote, the address of the visit, observations of the last visit.

    My doubts are the following:

    - My idea is to group the clients geographically, do I think this would be done using clustering?

    - and also assign the most qualified salesperson or, according to your sales level, visit a certain client. I think that you would use neural networks here?

    - Another question I have is whether the model would be one or I should use 2 models, one of clustering and another of neural networks. In case of using the 2, as would be done in Azure Machine Learning to process the 2 or only one model should be processed.

    I would appreciate if you can guide me or in any case if you can recommend a book of azure machine learning where you can explain how to work with several models at the same time.

    I say goodbye by thanking you in advance for your prompt answers and suggestions to my doubts.
    Friday, October 5, 2018 12:25 AM

All replies

  • Hi, Jaime,

    I am afraid your questions cannot be easily without additional context. I think it will be beneficial if you start with some basic training on Machine Learning before tackling the problem you proposed above. Andrew Ng's machine learning class on Coursera would be an excellent start. Once you grasped the basic concept of machine learning, you will be able to define the problem better and perhaps we can advise on whether clustering or neural network are viable approaches, and how Azure Machine Learning specifically can assist here. 


    - Hai

    • Proposed as answer by Hai Ning Friday, October 5, 2018 4:48 PM
    Friday, October 5, 2018 4:47 PM