RapidMiner has announced its first application for RapidMiner AI Cloud. RapidMiner Auto Cloud is an extension to RapidMiner Studio and assists with the process of building and validating models. Where Auto Model is impressive is that while it automates the analytics process for users it also enables them to understand exactly how it did it. This is different from many other modern solutions. These look faster and more efficient but you never quite know how they work. Auto Model is the first application for RapidMiner AI Cloud, more will follow. The same can be said for the modern car engine. Often they are sealed units with little way of understanding what actually lies beneath the cover. That is often the problem with modern technology. However, RapidMiner builds its solutions for data scientists who need to understand how the modelling process works.
What is Auto Model?
Auto Model is designed to address three task classes of problems that analysts face with an automated process and GUI interface. It assumes that the data is ready for the Auto Model process. A recent article highlights the process that users will run through with the new tool. The Auto Model takes the user through a six stage process to reveal insights in data that would have taken traditional methods far longer. The processes are:
- Select data
- Select Task: (These are the classes tasks that Auto Model addresses: Prediction, Clustering or Outliers).
- Prepare Target: (Deals with the classification of data. It assists the user by displaying data points for each class available up to a maximum of 10 classes.
- Select Inputs: The user selects which columns of data to use in the modelling. The tool delivers a value for each classes selected across four different pattern types: (Correlation, IDness, Stability and Missing data).
- Model Type: Various model types are supported depending in part upon the selections already made.
- Results: This displays results in a graphical dashboard that allows users to quickly modify or change the model.
Importantly, users are then able to drill down into the analytics tools used. The user is able to create scenarios quickly using the model simulator. They are also able to look at the processes in detail behind the automated model using a process viewer. This graphically lays out the steps taken to arrive at the selected results. It allows users to validate their choices and makes it easier to spot why their results did not come out as expected. This ability to look under the hood of the analytics tool is useful and sometimes missing from other solutions.
But my data isn’t ready
Fear not, for RapidMiner already has a tool to help with data preparation, Turbo Prep. It has also announced that three of the future enhancements for the AI cloud are around data and its preparation:
- An intuitive data prep application to transform, pivot and blend data from multiple sources with just a few clicks.
- Data enrichment services to automatically search and merge in additional data for better model generation.
- Integrations with leading data visualization and data integration platforms.
No details are available yet on when these will become available though.
What does this mean?
In addition, other enhancements are promised, with a real-time scoring engine for prescriptive analytics and some native collaboration features. RapidMiner continues to deliver innovation on its analytics platform. This was recognised by Gartner who placed it in the leader quadrant for data science and machine learning platform, for the fifth year running.
Lars Bauerle, Chief Product Officer at RapidMiner: “We’re thrilled to announce Auto Model on the RapidMiner AI Cloud. AI Cloud is built on a modern, powerful architecture that will scale to the most demanding customer use cases. With the release today of Auto Model for AI Cloud, anyone can get started building predictive models in just a few minutes. We have an exciting vision for the future of AI Cloud, and we will share more over the coming months.”
More companies are looking to move their analytics to the cloud. RapidMiner, building out its new set of functionality on the AI Cloud, is a positive move. It is a tool that appeals to data scientists. This latest application allows those data scientists to work faster whilst still applying their deep domain knowledge to understand the underlying methodology used.
