Back to GalleryNeural Network (MLP)
Multi-Layer Perceptrons are feedforward neural networks. By combining weights and non-linear activation functions across multiple layers, they can model highly complex, curved decision boundaries. They learn via backpropagation.
Key Parameters & Visual Influence:
• Hidden Layers: Adding layers/neurons increases capacity to model non-linear boundaries.
• Activation: Diverse functions (ReLU, Sigmoid, Tanh, Leaky ReLU, ELU, Mish, etc.) shape boundary curvature.
• Optimizer: Adam and RMSprop adjust learning rates dynamically to converge faster than SGD.
• Dropout: Randomly disables neurons during steps to prevent co-adaptation and overfitting.
Dataset Studio
CSV format: Headers in first row, last column = label/target. All values must be numeric.
Example: x1, x2, label