About The App
Discover MultilabelDTree, an automated machine learning web application for classifying multi-label data through decision trees. Our application incorporates cutting-edge techniques, including BinaryRelevance, LabelPowerset and ClassifierChain transformation approaches, to ensure that your data is classified accurately and efficiently. By integrating AutoML, we automate the entire process, allowing you to focus on extracting insights and making impactful decisions with the best accuracy.
BinaryRelevance
Transforms a multi-label classification problem with L labels into L single-label separate binary classification problems, and using a Decision Tree for each label.
LabelPowerset
Transforms the multi-label problem into a multi-class problem, by training a single Decision Tree classifier on all unique label combinations found in the training data.
ClassifierChain
Transforms a multi-label problem into a multi-class problem where each label is predicted sequentially, by using a Decision Tree for each label based on the Bayesian Chain rule.
How The App Works?
Log In or Sign Up to Unlock Full Access
Log in to your existing account, or sign up to experience our application features. Build new models and classify data using your top-performing pretrained models, all within a comprehensive and user-friendly environment.
Test Parameters & Evaluate Models
Test your parameters by choosing between pre-uploaded datasets or the ones you personally own and get your model evaluation. For even greater efficiency, enable Auto mode to let the app automatically discover the optimal values with the best accuracy.
Save Your Model & Visualize DTrees
Once the evaluation process is complete, save your highest-performing models in your account for future use. Easily export your models in PKL format, or visualize the Decision Tree graphs to gain deeper insights into your model's structure.
Classify Data using Pretrained Models
Finally, upload unclassified datasets and predict their labels by using your pretrained models. Once the predictions are complete, export the full results as a .csv file for easy analysis. Additionally, access detailed metrics to assess the quality and accuracy of your predictions.
