I last got an update on SPSS early last year. In March 2015, IBM released an update for the complete set of products in its advanced analytical portfolio. Analytical Decision Management (ADM) is positioned as part of this portfolio of predictive and prescriptive analytic tools.
IBM’s advanced analytic portfolio is focused on having meaningful impact on the “business moment” – actually changing the way decisions are being made to take a more impactful action. This focus on decision making, on taking action, is what makes ADM a key component. In addition they see simplifying analytics to reduce both the skills burden AND the time to insight/value. This means simplifying the engagement of different kinds of users, simplifying consumption through interactivity, simplifying actionable insight across data sources and simplifying deployment. The portfolio enables a set of analytic solutions. It is in itself a set of platform capabilities – Data Collection, Statistics, Modeler, ADM (included in IBM SPSS Modeler Gold) and Analytic Server. These capabilities are supported by IBM SPSS Collaboration and Deployment Services and complemented by additional technologies like IBM Decision Optimization, Watson Analytics, InfoSphere Streams and Social Media Analytics.
The 2015 Predictive Portfolio release included 3 main categories of new functionality:
- New algorithms built for scale
- Geospatial data sources and better integration of location and time data
- Richer deployment, especially on Bluemix and other cloud deployments – predictive analytics everywhere
Specific new functionality relevant to ADM includes:
- New algorithms in SPSS Modeler allow geospatial data to be easily prepared and integrated into models with support for geofences, spatial-temporal prediction, etc.
- Integration of R and Python as nodes in SPSS Modeler (previous release). There is also an ability to develop a custom user interface for a node so that new or specialized algorithms can be wrapped in a custom interface.
- Several algorithms were added or updated to include geospatial and time data and Big Data (Apriori/Association, Temporal Causal Modeling, Spatio-Temporal Prediction, TwoStep Cluster).
- SPSS Modeler is now connected to the CPLEX Optimization Studio so that, for instance, a modeler stream can be fed directly into an optimization model. This is in addition to the use of CPLEX in some of the analytic solutions. Support for optimization deployment to the cloud is in beta too.
- Analytical Decision Management has been updated to include improved user interfaces and capabilities especially around data preparation functions, reporting on projects, a specific user interface for defining and managing A/B or Champion/Challenger testing.
- An API for embedding predictive scoring (modeler streams) into other applications. These services can also be deployed on Bluemix.
- In addition to the Bluemix deployment the whole stack is available in the cloud through Softlayer.
IBM is one of the vendors in our Decision Management Systems Platform Technologies Report and there is more information on the portfolio here.