RapidMiner Studio is a comprehensive data science platform with visual workflow design and full automation
Features
Visual Workflow Designer
Increase productivity across the entire data science team, from analysts to experts
Connect to Any Data Source
Work with all of your data, no matter where it lives
Automated In-Database Processing
Run data prep and ETL inside databases to keep your data optimized for advanced analytics
Data Visualization & Exploration
Evaluate data health, completeness, and quality
Data Prep & Blending
Eliminate the hassle of preparing data for predictive modeling
Visual & Automated Machine Learning
Quickly create impactful machine learning models without writing code
Model Validation
Understand the true performance of a model before deploying to production
Explainable Models Not Black Boxes
Create visual data science workflows that are easy to explain and easy to understand
Get More From R & Python Code
Scalable code deployment and collaboration between coders and non-coders
Flexible Scoring & Model Operations
Turn predictive insights into business impact
Automation & Process Control
Build sophisticated visual workflows and automate important tasks
Open & Extensible
Integrate with existing applications and code
What’s New
Version 9.10.0:
Features:
- Added Function Fitting operator that can optimize parameters in a function of the attributes to fit the label. It can be used to create an optimal function to fit the data points in your data.
- Bias Awareness: if the use of a specific column is more likely to add unwanted bias to your models, it is highlighted as such. This happens in various places such as in the Statistics view of data, the model simulator, in Turbo Prep, in Auto Model, during model training, in model annotations among others.
Enhancements:
- The De-Normalization operator has a new parameter to also de-normalize predictions.
- Based on attribute name: prediction(abc) tries to use de-normalization of abc if no explicit de-normalization available
- The label (or other special attributes) can be included in normalization already in the normalize operator. The changes allow for multiple prediction attributes to be affected
- Added date format parameter to Write CSV in case format date attributes is selected
- Improved performance of Append operator
- Handled yet another case of JDBC drivers ignoring the JDBC standard gracefully (here: Infor Data Lake DatabaseMetaData#getTypeInfo())
- Introduced operator signatures to improve the startup of Studio
- Signatures contain meta information that is used in operator registration, global search setup and documentation browser display
- Signatures are persisted between starts for an improved startup time
- Signature persistence can be configured or cleared with the setting System -> Local File Cache -> Keep Operator Signatures
- Time Series: Enabled the usage of constant values for the replace types in the Equalize Numerical Indices and Equalize Time Stamps operators
- The operators can now be used to fill gaps in non-equal data sets with constant values
- Time Series: All Time Series operators (except for Multi Horizon Forecast, Multi Horizon Performance) now working with Belt IOTable (as in- and output)
Bugfixes:
- In rare instances, operator parameters did not get saved correctly if a default value was set for it. This e.g. affected date parameters used in extensions.
- Generate Attributes max and min functions do now always return missing value if any of the values is missing.
- Fixed missing operator help for Azure Blob Storage and Data Lake Storage operators
Compatibility
OS X 10.8.0 or later
Screenshots
Download Now
