20 algorithms · 13 distributions · all interactive

See the maths move.

Infera turns machine learning and statistics into things you can poke. Drop points, drag them around, change a parameter and watch every model and distribution redraw in real time, with plain-English explanations and the formulas right beside it.

Explore algorithms Explore distributions

Play first

Every page opens on a live canvas. Click to add points, drag them, switch datasets, crank the noise.

Turn the knobs

Depth, kernels, learning rates, priors. Each control says what it does, and the picture answers instantly.

Then read why

Underneath: the intuition, the step-by-step algorithm, the formulas, and experiments worth trying.

Algorithms

See all

Regression

Predict a number. Fit a curve through points and see how flexible each model is.

Linear Regression

Least squares through your points, from a straight line to a wiggly polynomial.

Regression

KNN Regression

Predict the average of the k closest points. No equation, just neighbours.

Regression

Regression Tree

Split the x-axis into steps and predict the mean inside each one.

Regression

Gradient Boosting

Add small trees one at a time, each one fixing what the others got wrong.

Regression

Linear classifiers

Separate classes with a line, or with a line in a richer feature space.

Logistic Regression

A linear score squashed into a probability, trained by gradient descent.

Classification

Perceptron

The original neuron. Nudge the line every time it gets a point wrong.

Classification

Probabilistic classifiers

Model each class as a Gaussian and let Bayes' rule pick the winner.

LDA

One Gaussian per class, all sharing the same shape, which gives linear boundaries.

Classification

QDA

Every class gets its own Gaussian shape, so boundaries become curves.

Classification

Naive Bayes

Assume the features are independent within each class, then apply Bayes' rule.

Classification

Instance-based

No training at all. Look at the neighbours and copy their answer.

KNN

Classify a point by a vote among its k closest neighbours.

Classification

Trees and ensembles

Ask yes/no questions about the features, then combine many trees into one strong model.

Decision Tree

A flowchart of yes/no questions that carves the plane into boxes.

Classification

Random Forest

Many slightly different trees vote, and the noise averages out.

Classification

AdaBoost

Weak learners in sequence, each focusing on the points the last one missed.

Classification

Kernels and neural networks

Bend the boundary with kernels or learn the features with a network.

SVM

The widest possible street between classes, bent by a kernel when needed.

Classification

Neural Network

Layers of neurons learn their own features, trained live with backpropagation.

Classification

Unsupervised

No labels. Find groups and directions hiding in the data.

K-Means

Drop k centres, assign points, move centres, repeat until nothing changes.

Clustering

Gaussian Mixture

Soft clustering with stretched ellipses, fitted by expectation-maximisation.

Clustering

DBSCAN

Clusters are dense regions. Any shape works, and outliers are left out as noise.

Clustering

Hierarchical

Merge the two closest groups, again and again, and read the clusters off a dendrogram.

Clustering

PCA

Find the directions where the data spreads out the most.

Dimensionality reduction

Distributions

See all

Bernoulli

X ~ Bernoulli(p)

A single yes/no trial that succeeds with probability p.

Binomial

X ~ Bin(n, p)

How many successes in n independent yes/no trials.

Geometric

X ~ Geom(p)

How many trials until the first success.

Poisson

X ~ Pois(λ)

How many events happen in a fixed window when they arrive at a steady average rate.

Discrete Uniform

X ~ U{a, b}

Every whole number from a to b is equally likely, like a fair die.

Uniform

X ~ U(a, b)

Any value between a and b is equally likely.

Normal (Gaussian)

X ~ N(μ, σ²)

The bell curve. Sums of many small effects end up looking like this.

Exponential

X ~ Exp(λ)

The waiting time until the next event in a Poisson process.