optimise parameters

k_vishnu772k_vishnu772 Member Posts: 34 Contributor I
edited November 2018 in Help

Hi All,

i have a small question regarding optimising the parameters for a model.

i would like to know is there any way the model can check the range of values for a specific parameters and give me the result based on that.

For example in titanic data set we  have age column and class (lower ,middle,upper),i fix the age range from 20 to 30 and fix the class value as lower ,i would like the see which age group from 20 to 30 in lower desk has high possibility of survival.

Is there a way i can optimise the model for the above case?





  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    @k_vishnu772 yes, you should use the Optimize Parameters operator for that. I do some optimization in my latest Live Stream video here: https://youtu.be/WdYpWAFxzR8


    It's about the 39 minute mark. 

  • k_vishnu772k_vishnu772 Member Posts: 34 Contributor I

    HI Thomas ,


    What i am looking is little different ,here i can only change the options for the operators in the model like no of trees or maximum depth , but i want to change the data in the attribute.i want to see like in titanic data set for age between 20 to 40 at what age the people has the highest survival rate so by fixing the age between 20 to 40 and i want see the parameters with highest survual rate.


    I came across this kind of feature in  Mando Brain Augumeneted intelligence tool ,so i am wondering is there NY similar kind of options in rapidminer.In model simulator i can only fix the all the values and see the result, but not like fixing some attribute  the values and changing the range of attribute values .




  • SGolbertSGolbert RapidMiner Certified Analyst, Member Posts: 344 Unicorn

    Hi Vishnu,


    Training a logistic regression model could help you analyze which features impact more upon the survival and even compare the relative risks of different realizations of the variables. It kind of sucks for forecasting, but it's good for interpretation.




Sign In or Register to comment.