The Altair Community is migrating to a new platform to provide a better experience for you. The RapidMiner Community will merge with the Altair Community at the same time. In preparation for the migration, both communities are on read-only mode from July 15th - July 24th, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here.
Options

No variance in confusion matrix output

LeoKrakLeoKrak Member Posts: 2 Newbie
edited May 2020 in Help
Hi, 

If anyone is available now, maybe i can go in further details but in brief my team and I are building a model that should predict who among a list of

 previous donators would donate for the next campaign, based on a database with previous donators and their characteristics.

We seem to do everything right but I think we face an overfitting issue because we always get the same results in the confusion matrix no matter how many tries we make.

Also, the model has extreme poor performances but what is really disturbing is that the output doesnt vary at all.. Any help ?

Answers

  • Options
    varunm1varunm1 Moderator, Member Posts: 1,207 Unicorn
    Hello @LeoKrak

    Your data sets seem to be highly imbalanced with more than 3700 samples belong to one class and just 54 to another. You need to try data preprocessing methods like sampling (on training side) or weighting. 

    Also, try to do some feature selection and tune your hyperparameters of the model, you can do so by using Optimize parameter (Grid) operator.
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

  • Options
    LeoKrakLeoKrak Member Posts: 2 Newbie
    Thanks a lot I'll try this tomorrow morning !  Can I come back to you if I have any other questions ?
  • Options
    varunm1varunm1 Moderator, Member Posts: 1,207 Unicorn
    Sure, we are here to help
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

Sign In or Register to comment.