Feature Request: Interactive Hyperparameter Tuning with Sliders & Model Performance Comparison #36
Description
Description:-
Currently, AlgoLab provides a way to select ML algorithms but lacks interactive hyperparameter tuning and direct model performance comparison.
This feature will allow learners to dynamically adjust algorithm parameters (e.g., k for KNN, tree depth for Decision Tree) and compare results visually in real time.
Proposed Features:-
Interactive Sliders for Hyperparameter Tuning
Add Streamlit sliders to adjust key hyperparameters:
KNN: Number of neighbors (k)
Decision Tree: Max depth, min samples split
Logistic Regression: Regularization strength
SVM: Kernel type, C value
Update decision boundary visualization dynamically as sliders change.
Model Performance Comparison
Allow selecting multiple algorithms (e.g., KNN vs Logistic Regression).
Train and display:
Accuracy, Precision, Recall, F1-Score in a comparison table.
Bar chart showing performance metrics side-by-side.
Tech Stack:-
Python
Streamlit → sliders, UI elements
scikit-learn → model training and metrics
Matplotlib / Plotly / Seaborn → visualization
Pandas, Numpy → data handling
Benefits:-
Enhances interactivity and engagement for learners.
Makes understanding the effect of hyperparameters visual and intuitive.
Helps learners compare algorithms to choose the best one for a dataset.
I would like to work on this as a GSSoC '25 contributor.
@manasvi-0, Please assign this issue to me if approved!