Synthetic Classifier Simulator
Every change retrains models and refreshes the confusion matrices.
Selected Traits
Save Run
Classes
Loading…
Welcome to the TBAC Playground
This playground is made available to to help understand the difference between classification accuracy and classification capacity in the case of ASD detection. With the TBAC Playground, you can simulate multiple synthetic datasets with up to 18 ASD traits and as many classes as you want (ASD, ADHD, ...) and manually set the characteristic. Observe how the confusion matrices and metrics evolve when you change the parameters of the data distribution and the training process.
About this project
TBAC playground is part of the research paper "TBAC18- a new framework to reconcile Machine Learning and practice for ASD detection" by Barry Martin et al. , here (Not Available Yet). The paper and the code for the playground are available on GitLab.
Playground walkthrough
Playground image placeholder 1
Short explanation for image 1.