I’m Sunny on the product marketing team at RapidMiner. We just launched RapidMiner 8.1, and I wanted to share some of my favorite things with you. Starting with one of the most exciting new features we’ve ever added to RapidMiner Studio!
RapidMiner Auto Model
The RapidMiner Auto Model automates machine learning and accelerates Data Science, making the platform more accessible to new users and more powerful for expert Data Scientists. It delivers new capabilities by transforming data to generate actionable insights, making your data science teams more productive, without compromising on transparency or control.
“The medical community would never accept a black box- the Auto Model rides through my data and modeling in a transparent, easy to follow and visually powered process.” – Sven van Poucke, MD, PhD
This new capability condenses a decade of experience in data science innovation at RapidMiner and is delivered to you in an easy to use, visually powered wizard, helping you get started with your data faster than ever before. It accelerates modeling by suggesting the best machine learning techniques, and automatically generates optimized, cross-validated models from your data.
The Auto Model will save you a lot of time in creating processes for multiple models. It shows which features have the greatest impact on your objective, with the most important influence factors and correlations. Built in visualizations and an interactive simulator let you explore the model to see how it performs. Best of all, there are no black boxes! Auto Model generates a RapidMiner Studio process behind the scenes so you fine tune, test, and quickly deploy the model into production.
The model simulator enables data scientists to gain a deeper understanding of the model, allowing them to easily communicate and validate their models with other team members and stakeholders. You can now optimize individual parameters and see the impact on your models and results, helping you find the optimal ‘next steps’ in your business process. The model simulator goes beyond predictive analytics, delivering data-driven prescriptive insights easily accessible to everyone in your organization.
This is a feature that has got a lot of data scientists excited and is going to make life easier for all our users, speeding up their daily work as global search sifts through their repositories to quickly retrieve anything of interest- including processes, models, operators, extensions and even UI actions. No need to search through your complete folder structure anymore: everything is now at hand!
We have re-factored a few operators, including Join, Correlation Matrix and K-Means to drastically improve performance, with up to x10 increases in speed.
Powerful operator toolbox
Use the new Replace Rare Values operator to automatically replace values of nominal attributes with low frequency. The new Merge operator merges Example Sets together by appending all attributes to one Example Set. And finally, we have made improvements to the Parametric Probability Estimator operator, which now allows the user to calculate the probability that a given value belongs to a class in a classification problem.