Machine Learning Studio (classic): Algorithm and module help
Support for Machine Learning Studio (classic) will end on 31 August 2024. We recommend you transition to Azure Machine Learning by that date.
Beginning 1 December 2021, you will not be able to create new Machine Learning Studio (classic) resources. Through 31 August 2024, you can continue to use the existing Machine Learning Studio (classic) resources.
- See information on moving machine learning projects from ML Studio (classic) to Azure Machine Learning.
- Learn more about Azure Machine Learning.
ML Studio (classic) documentation is being retired and may not be updated in the future.
Customers currently using or evaluating Machine Learning Studio (classic) are encouraged to try Azure Machine Learning designer, which provides drag-n-drop ML modules plus scalability, version control, and enterprise security.
Machine Learning Studio (classic) is a cloud predictive analytics service that makes it possible to quickly create and deploy predictive models as analytics solutions. The machine learning tools are mostly cloud-based services, which eliminates setup and installation concerns because you can work through your web browser on any internet-connected PC. See the article, "What is Studio (classic)?" for more details.
This documentation contains detailed technical and how-to information for the modules that are available in Machine Learning Studio (classic).
- Sign in to your Machine Learning Studio (classic) workspace and get started.
What is a module?
Each module in Machine Learning Studio (classic) represents a set of code that can run independently and perform a machine learning task, given the required inputs. A module might contain a particular algorithm, or perform a task that is important in machine learning, such as missing value replacement, or statistical analysis.
In Studio (classic), modules are organized by functionality:
Data input and output modules do the work of moving data from cloud sources into your experiment. You can write your results or intermediate data to Azure Storage, a SQL database, or Hive, while running an experiment, or use cloud storage to exchange data between experiments.
Data transformation modules support operations on data that are unique to machine learning, such as normalizing or binning data, feature selection, and dimensionality reduction.
Machine learning algorithms, such as clustering, support vector machine, or neural networks, are available within individual modules that let you customize the machine learning task with appropriate parameters. For classification tasks, you can choose from binary or multiclass algorithms.
After you've configured the model, use a training module to run data through the algorithm, and measure the accuracy of the trained model by using one of the evaluation modules. To get predictions from the model you've just trained, use one of the scoring modules.
Anomaly detection: Machine Learning Studio (classic) includes multiple algorithms specialized for these tasks.
Text analytics modules support various natural language processing tasks.
Vowpal Wabbit support makes it easy to use this scalable platform.
OpenCV library provides modules to use in specific image recognition tasks.
Time series analysis supports anomaly detection in time series.
Statistical modules provide a wide variety of numerical methods related to data science. Look in this group for correlation methods, data summaries, and statistical and math operations.
In this reference section, you'll find technical background on the machine learning algorithms, implementation details if available, and links to sample experiments that demonstrate how the module is used. You can download examples in the Azure AI Gallery to your workspace. These examples are for public use.
If you are signed in to Machine Learning Studio (classic) and have created an experiment, you can get information about a specific module. Select the module, then select the more help link in the Quick Help pane.
Other technical references
|Data Types List
|This section contains reference topics describing the learner interfaces, and the
DataTable format used for datasets.
|This section lists the errors that modules can generate, with causes and possible workarounds.
For the list of error codes related to the web service API, see Machine Learning REST API error codes.