Automated machine learning
Automatically build machine learning models with speed and scale
Easily build highly accurate machine learning models
Empower professional and non-professional data scientists to build machine learning models rapidly. Automate time-consuming and iterative tasks of model development using breakthrough research – and accelerate time to market.
Automatically build and deploy predictive models using the no-code UI or using the SDK
Rapidly create accurate models customised to your data and refined by a wide array of algorithms and hyperparameters
Increase productivity with easy data exploration and intelligent feature engineering using deep neural networks
Build responsible ML solutions with model interpretability, and fine-tune your models to improve accuracy
Accurately forecast future business outcomes with popular time series models and deep learning
Build models your way
Accelerate model creation with the Automated machine learning no-code UI, or SDK. Customise your models quickly and apply control settings to iterations, thresholds, validations, blocked algorithms and other experiment criteria. Use built-in capabilities for common machine learning tasks such as classification, regression and time-series forecasting to handle large datasets and improve model scores.
Control the model building process
Automated machine learning intelligently selects from a wide array of algorithms and hyperparameters to help build highly accurate models. Discover common errors and inconsistencies in your data through guardrails, and better understand recommended actions and apply them automatically. Use intelligent stopping to save time on compute and prioritise the primary metric to speed results.
Improve productivity with automatic feature engineering
Use built-in capabilities for common machine learning tasks such as classification, regression and time-series forecasting, including deep neural network support, to handle large datasets and improve model scores. Utilise the automated feature selection and new feature generation capabilities to save time and build highly accurate models. Automated ML now includes BERT deep learning architecture for text data featurisation in 100 languages, which is available through the UI as well as the Notebooks.
Understand models better
Built-in support for experiment run-summaries and detailed metrics visualisations help you understand models and compare model performance. Model interpretability helps evaluate model fit for raw and engineered features and provides insights into feature importance. Discover patterns, perform what-if analyses and develop deeper understanding of models to support transparency and trust in your business.
Support a variety of machine learning tasks
Get support for essential machine learning applications such as classification, regression and time-series forecasting, including special built-in featurisers to configure each task. Use classification techniques for supervised learning, where common applications include fraud detection, handwriting recognition etc. Build regression models to predict numerical values such as price prediction. Or use time-series forecasting to build models that consider trends and seasonality. Evaluate a variety of popular time-series models including ARIMA and Prophet and other deep learning models.
Get started with automated machine learning
See how you can accelerate AI at scaleView an automated machine learning infographic
Machine learning concepts
Build your machine learning skills with Azure
Learn more about machine learning on Azure and participate in hands-on tutorials with this 30-day learning journey. At the end of this learning journey, you’ll be prepared to take the Azure Data Scientist Associate certification.