Which machine learning and AI algorithms are covered?

Linear Regression, Logistic Regression, Decision Tree (CART), Support Vector Machine (SVM), k-Nearest Neighbors, Gaussian Naive Bayes, Multi-Layer Perceptron (Neural Network), k-Means Clustering, Fuzzy C-Means (FCM), Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Q-Learning Gridworld, and Feature Normalization.

Can I use my own data, or only presets?

You click directly on the 2D plot to place your own data points (and, for classification, assign them to Class A or Class B), then run the algorithm on exactly the data you plotted.

Is any computation or data sent to a server?

No. Every gradient descent step, distance calculation, backpropagation pass, swarm velocity update, and Bellman equation executes 100% locally in your browser memory.

Does this show Scikit-learn Python code?

Yes. Every topic pairs its interactive animation with mathematical formulas, step pseudocode, and real-world Python Scikit-learn / scientific code snippets.

Machine Learning Visualizer NEW

Click on the plot to place your own data points, then run the algorithm and watch it converge step by step, with mathematical formulas, pseudocode, and production Scikit-learn references.

Private by architecture: All processing is 100% client-side memory. Your data never leaves this browser tab.

Click the plot to add points, then press Run.

Step 0 / 0

Mathematical Formula

ŷ = wx + b

Predict y as a linear function of x. Fit w and b by minimizing mean squared error.

Execution Algorithm

Scikit-Learn Reference

from sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(X, y)

Production equivalent in Python scikit-learn.

About This Tool

Machine learning algorithms are usually taught through static equations and pre-rendered textbook plots. This visualizer lets you plot your own 2D data by clicking, then watch decision boundaries form, margins optimize, swarms converge, and neural networks train, step by step.

Spans the spectrum of modern intelligent systems: classic regression, classification trees, maximum-margin SVMs, backpropagation in deep neural networks, unsupervised clustering and fuzzy memberships, swarm and evolutionary optimization, and tabular reinforcement learning.

Key Features

Click-to-Plot Data: Place your own points directly on the 2D canvas instead of relying on fixed presets.
13 Fundamental ML & AI Topics: Spans Regression, Classification, Deep Learning, Clustering, Fuzzy Logic, Evolutionary Optimization, and RL.
Scikit-Learn Python References: Every algorithm includes production-ready Python code snippets matching official scikit-learn conventions.
Formulas & Step Breakdown: Real mathematical notation (Gini, Margin, Bellman, Softmax, Sigmoid) with plain-language step explanations.
100% In-Browser Memory: Zero tracking, zero server dependencies, and immediate client execution.

Frequently Asked Questions

How do Decision Trees choose axis-aligned split points?

The tree evaluates candidate thresholds along feature axes and selects the split that maximizes Gini reduction (or Information Gain), minimizing class impurity in child nodes.

What are Support Vectors in an SVM?

Support vectors are the critical data points that lie directly on or inside the margin gutters. They dictate the orientation and width of the maximum-margin hyperplane.

How does Fuzzy C-Means differ from crisp k-Means?

In standard k-Means, points are assigned strictly to 1 cluster. In Fuzzy C-Means, each point possesses a continuous degree of membership across all clusters, showing overlap and uncertainty.

How does Particle Swarm Optimization avoid local minima?

Particles balance inertia with two gravitational pulls: cognitive memory (personal best position) and social cooperation (global swarm best position), exploring the landscape collaboratively.

How does Q-learning discover an optimal policy?

Using the Bellman optimality backup equation, the agent updates expected state-action values Q(s, a) from environmental step rewards and discounted future value estimates.