Big Data Predictive Analytics Solutions Wave 2015
Recent advances in predictive and prescriptive data science techniques is providing savvy organizations with competitive advantage.
New improvements in predictive analytics tech includes the use of APIs and Predictive Modeling Markup Language (PMML); the integration of predictive analytics into business apps; and the availability of “one-click predictive modeling” applications and marketplaces for predictive apps.
Leaders
IBM offers one of the most complete set of capabilities for building and deploying predictive models, either on-prem or in the cloud.
SAS offers a nice set of tools that integrate with R, Python, and Hadoop.
SAP is moving up in the analytics world with its HANA in-memory technology and PAL, or Predictive Analytics Library.
Strong Performers
RapidMiner offers a “rock solid” enterprise solution with more than 1,500 “methods” that address all stages of the analytics lifecycle and has among the tightest integration with the cloud.
Alteryx is particularly strong in the data preparation component of the lifecycle, and also impressed with its apps gallery.
Oracle has built tight integration between its database and analytical workflows built with its SQL Developer tool.
FICO offers its “incredibly deep knowledge” for general purpose predictive modeling.
Dell offers a “comprehensive library” of algorithms and tools (thanks to its acquisition of Statistica).
Angoss may be a good bet for a company just starting down their predictive analytics journey, thanks to its “intuitive interface” and continued leadership in decision trees.
Alpine Data Labs offers a comprehensive collection of tools, with ability to push algorithms down into Hadoop.
KNIME is an open source software developer that delivers “robust” modeling tools that are as good as the proprietary vendors.
Contenders
Microsoft with the Azure Machine Learning cloud is a bit green, but its acquisition of Revolution Analytics should help.
Predixion Software has an Excel-based offering that allows data scientists “to effortlessly build” models in the cloud. Predixion’s machine learning semantic model (MLSM), which packages models up in.NET and Java containers is cool.