Global AI and Data Science

 View Only

  • 1.  Webcast Into Data Science: Understanding Decision Trees Follow-up

    Posted Fri March 22, 2019 01:00 PM

    Thank you everyone for attending today's webcast! In this webcast we covered decision trees:

    • What are they?
    • What are they good for?
    • Decision trees limitations
    • Types of decision trees
    • A demo showing the use of decision trees for customer churn prediction

    The replay can be found here. If you have any questions, please feel free to post them in this thread.

    ​

    ------------------------------
    Jacques Roy
    ------------------------------

    #Askadatascientist
    #GlobalAIandDataScience
    #GlobalDataScience


  • 2.  RE: Webcast Into Data Science: Understanding Decision Trees Follow-up

    Posted Fri March 22, 2019 01:04 PM
    Hi Jacques,

    In the webinar, you said I reckon that decision trees are a simple form of ML. What if decision trees are not sufficient, how should one proceed if that's the case?

    Thanks!

    ------------------------------
    Henri Ajenstat
    ------------------------------

    Original Message


  • 3.  RE: Webcast Into Data Science: Understanding Decision Trees Follow-up

    Posted Fri March 22, 2019 03:47 PM
    When I said that decision trees were simple, I did not mean it is a simple form of ML. I meant that they are easily understood. They are also transparent in the sense that we can figure out how we got to a decision.
    When you ask about the situation where decision trees are not sufficient, you probably means that the accuracy is not high enough.
    There could be many reasons:
    • Data quality
    • Not enough data / data not representative of the problem domain
    • Too many attributes
    • Having a tree that overfits
    and probably more.
    I did mention in the webcast that there are other algorithms that use decision trees but help in the accuracy. For example, random forests and gradient boosted trees. This could help in getting better accuracy.
    of course, there are multiple ways to solve one problem so other approaches could be taken, including neural network. Still, don't give up on decision trees too fast :-)

    ------------------------------
    Jacques Roy
    ------------------------------

    Original Message


  • 4.  RE: Webcast Into Data Science: Understanding Decision Trees Follow-up

    Posted Fri March 22, 2019 01:31 PM
    Hello Jacques,

    Ensembles need not necessarily be all decision trees; could be a decision tree and different algorithm(s), right?

    Thank you!

    ------------------------------
    Rajagopal Krishnarjunan
    ------------------------------

    Original Message


  • 5.  RE: Webcast Into Data Science: Understanding Decision Trees Follow-up

    Posted Fri March 22, 2019 03:52 PM
    Definitively it involves more than just decision trees.
    Take for example random forests. They divide they attributes randomly between multiple trees and then another algorithm figures out how to weigh each result to arrive at an overall decision.
    The nice thing about this is that you only need to be aware that this happens instead of having to implement it yourself. Then you can look at different type of ensembles and figure out which one is more appropriate to your problem.

    ------------------------------
    Jacques Roy
    ------------------------------

    Original Message


  • 6.  RE: Webcast Into Data Science: Understanding Decision Trees Follow-up

    Posted Fri March 22, 2019 04:05 PM

    Thanks Jack. Looking forward to attending the future sessions of this series.



    This message is for the designated recipient only and may contain privileged, proprietary, or otherwise confidential information. If you have received it in error, please notify the sender immediately and delete the original. Any other use of the e-mail by you is prohibited. Where allowed by local law, electronic communications with Accenture and its affiliates, including e-mail and instant messaging (including content), may be scanned by our systems for the purposes of information security and assessment of internal compliance with Accenture policy. Your privacy is important to us. Accenture uses your personal data only in compliance with data protection laws. For further information on how Accenture processes your personal data, please see our privacy statement at https://www.accenture.com/us-en/privacy-policy.
    ______________________________________________________________________________________

    www.accenture.com



    Original Message