What Is Statistics and Machine Learning Toolbox?

Statistics and Machine Learning Toolbox™ provides functions and interactive apps to describe, analyze, and model data in MATLAB^®. Use descriptive statistics and visualizations to explore data, fit probability distributions, perform hypothesis tests, generate random samples for Monte Carlo simulations, and quantify variation to support data-driven decisions.

For predictive modeling, use correlation analysis, feature selection, principal component analysis, and dimensionality reduction to identify informative predictors and prepare tabular data. The toolbox supports supervised, semi-supervised, and unsupervised classical machine learning for regression, classification, clustering, and anomaly detection. In Classification Learner and Regression Learner, train, tune, compare, validate, and interpret models interactively, then move to MATLAB code without rebuilding the workflow. Export models and results, generate reusable MATLAB code or machine learning pipelines, and apply the analysis to new data.

The toolbox also supports tools for industrial statistics, including design of experiments, survival and reliability analysis, statistical process control, and measurement system analysis. DOE Explorer and Gage R&R Analyzer guide specific quality-engineering workflows and export analyses to the MATLAB workspace.

Integrate supported trained models into Simulink^® for simulation and Model-Based Design. With App Designer, build custom apps around workflows specific to your team and share them across the organization. Generate C or C++ prediction code with MATLAB Coder™, or package MATLAB apps and analytics for standalone and enterprise deployment with MATLAB Compiler™.