How i could use RapidMiner for search in the web for prove my teorical model?

AlandeyAlandey Member Posts: 8 Contributor II
edited December 2018 in Help

Using Data Miner, I need to prove that a construction of my Theoretical Model has a strong correlation. How could I use RapidMiner to search the web to prove, for example, considering smart cities and IoT, that a construct of my theoretical model, namely Logistics, has a strong relationship with sustainability?

Answers

  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member, University Professor Posts: 568 Unicorn

    Hello, @Alandey!

     

    Your question is a bit too broad, I guess. Let me give you an example, so you can continue preparing your data with this in mind. I have a data set with the following data:

     

    • Systolic.
    • Diastolic.

    Systolic and Diastolic blood pressures are two numerical variables that might or might not be correlated (in my case, they are). To prove that there is a correlation there are many methods. You can, for example, build a scatter plot as a visualization. Here is one:

     

    Screen Shot 2018-09-07 at 01.12.11.png




    Before measuring two seemingly different numerical values (diastolic blood pressure is always lower than systolic), I Normalized the numbers and wrote these. Notice how you can trace a diagonal line and see that the points are not that far from that line.

     

    Another graphical method to show that there is a correlation is to use a series chart. See how the diastolic isn't that far from systolic blood pressure.

     

    Screen Shot 2018-09-07 at 01.15.18.png

    Now, for the ultimate algorithms you can use to demonstrate a strong correlation between two variables, those would be the correlation matrix and the linear/logistic regression. The first one is easier.

     

    Screen Shot 2018-09-07 at 01.18.28.png

     

    There are linear regressions, logistic regressions and other stuff that can allow you do that. All you have to do is to convert your data to numerical to perform the analysis.

     

    Hope this helps, but as I told you, your question seemed a bit too open to me :(

     

    All the best,

     




  • AlandeyAlandey Member Posts: 8 Contributor II

    Thank rfuentealba

    And in the case that i dont have the date, in other words, my date saw mining and need reseach in the internet? How to use the Rapidminer for make this research categorinzing whatever for Logistic and Sustentability and next view the relation ?

     

  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member, University Professor Posts: 568 Unicorn
    I think we would need more details on your kind of research.

    If you can, send your data and processes. If you can't, send some data in a way that we can understand what are you trying to accomplish. Sorry for not being more helpful.

    All the best,
Sign In or Register to comment.