Machine learning is an elemental core competency. It is a fundamental building block to AI. It gives enterprises the power to predict. Most importantly, it can make enterprises gain the agility of disruptive upstarts by injecting scalable intelligence into customer experiences, business process applications, and employee decisions. That’s where predictive analytics and machine learning (PAML) solutions come in to help data science teams crank out machine learning models and to collaborate with the rest of the organization to use them in production applications. In two Forrester Waves™, we have evaluated a total of 22 vendor products.
Three Types Of PAML Solutions — Two Forrester Waves™
Forrester recognizes three distinct market segments for machine learning solutions that we identified in an earlier report. We have published two Waves, with a third planned next year to cover automation-focused vendors.
Multimodal PAML solutions provide the widest breadth of workbench tools. These solutions offer multiple user-interface paradigms and the broadest set of workbench tools, such as graphical user interfaces (GUIs), configuration wizards, automation, and coding environments. Many of these solutions also provide tools for non-data scientists to build data pipelines, create machine learning models, and collaborate with data science teams. Forrester clients may read the report here.
Vendors evaluated in the Wave: Dataiku, Datawatch, FICO, IBM, KNIME, MathWorks, Microsoft, RapidMiner, Salford Systems (Minitab), SAP, SAS, TIBCO Software, and World Programming
Notebook-based PAML solutions favor a code-first approach. Notebook-based PAML solutions provide workbench tools centered on coding in R, Python, and other programming languages using open source Jupyter or a proprietary interface that makes coding more efficient. The vendors in this segment add significant, differentiated features, such as environment provisioning, project management, deployment, model management, visualization tools, and more. Forrester clients may read the report here.
Vendors evaluated in the Wave: Anaconda, Civis Analytics, Cloudera, Databricks, Domino Data Lab, Google, H2O.ai, OpenText, and Oracle
Automation-Focused PAML Wave Planned For 2019
Automation-focused PAML solutions help non-data scientists build models. This segment focuses on tools to automate the steps in the model-building life cycle. Automation-focused solutions enable data scientists and non-data scientists to build models by configuration, instead of coding and specifying each step in a data science pipeline. Some multimodal and notebook-based vendors offer automation as well, but they also offer other approaches to building models and thus are not exclusively included in this automation-focused segment.