Rapid-I, a leading provider of open source solutions for predictive analytics, data mining and text mining, is releasing three extensions for its solutions RapidMiner and RapidAnalytics. The extensions optimize the exchange of data mining results with other applications via PMML and support the connection with the web as a source of data. Following the Wikipedia idea, the transfer of user-generated content (UGC) into the world of data mining is also made possible by a new community extension. "RapidMiner is the number 1 open source solution worldwide for process-based data analyses. Broadening the range of functions, which is already wide as it is, by way of so-called extensions has been part of our overall strategy since the beginning", explains Dr. Ingo Mierswa, CEO of Rapid-I.

PMML Extension

An important factor for the integration of data mining results into other applications is the Predictive Model Markup Language (PMML). This industry standard serves for the exchange of data mining models between different applications and platforms. Models which are produced with RapidMiner can now be directly transferred to a database which also supports the PMML standard and thus be easily led to deployment. Apart from the exchangeability of data mining models, the possibility of applying models to large data sets directly in the database is an especially important topic for Rapid-I. Until now, RapidMiner users had to rely on combinations of special access operators for selected modelling methods like Naive Bayes in order to ensure scalability for large volumes of data which is now no longer necessary.

"With the new PMML Extension for RapidMiner and RapidAnalytics, the possibilities of using RapidMiner models for scoring large data sets have become greater, thus making the deployment of data mining a whole lot easier", says Mierswa.

Web Extension

The new web extension for RapidMiner and RapidAnalytics is available for a stronger integration with information from the Internet. This extension is a complete revision of access methods from the text extension of the older versions. The web extension makes it possible to access all information from the web as well as combine them with additional information. Additionally to a built-in web-crawler, which serves for collecting contents from the Internet for data analyses, the extension supports specific functions for working with webtexts as well as for docking onto other Internet sources such as RSS feeds. This means that enterprises can no longer only access their structured information in databases, but consult any data from the Internet for evaluation and thus increase their knowledge substantially.

Community Extension

With the new community extension, RapidMiner is bringing the basic idea of the Web 2.0 into the data mining world. The extension offers a direct connection to the portal myExperiment.org, where users can discuss their data mining processes. In addition, it is possible to communicate with data analysts who are working on similar analysis problems. The portal myExperiment.org already has an active community and also has all the characteristics of a social network. The new extension makes uploading processes possible with one click as well as searching for your own processes and even those of other users if they are shared. These processes can then be very easily downloaded and applied to your own data - again with a single click.

"The community extension for the connection to myExperiment.org is a great addition to the new data and process repositories from RapidMiner 5.0 and is strongly based on the very successful idea of Wikipedia", notes Dr. Simon Fischer, Head of Research and Development at Rapid-I. "Based on these workflows, Rapid-I will also develop further tools for supporting users in the design of new processes."

Links