Modeler can read data from flat files, spreadsheets, major relational databases, IBM Planning Analytics and Hadoop. Extend Modeler's capabilities to push back data processing with the SQL Optimization add-on (subscription) or the Analytic Server (perpetual licenses).
Visual analysis streams
Use an intuitive graphical interface to visualize each step in the data mining process as part of a stream. Analysts and business users can easily add expertise and business knowledge to the process.
Automatic data preparation
Transform data automatically into the best format for the most accurate predictive models. Analyze data, identify fixes, screen out fields and derive new attributes with just a few clicks.
Use a single run to test multiple modeling methods, compare results and select which model to deploy. Quickly choose the best performing algorithm based on model performance.
A range of algorithmic methods
Choose from multiple machine learning techniques, including classification, segmentation and association algorithms including out of the box algorithms leveraging Python and Spark. Use languages such as R and Python to extend modeling capabilities.
Capture key concepts, themes, sentiments and trends by analyzing unstructured text data. Uncover insights in blog content, customer feedback, emails and social media comments.
Explore geographic data such as latitude and longitude, postal codes and addresses. Combine it with current and historical data for better insights and predictive accuracy.
Support for open source technologies
Use R, Python, Spark and Hadoop to amplify the power of your analytics. Extend and complement these technologies for more advanced analytics while you maintain control.
Multiple deployment methods
Using Modeler Gold, data scientists can schedule jobs to run at desired times. IT administrators can integrate deployment into existing systems for batch, real-time or streaming (through IBM Streams). Customers can deploy SPSS Modeler programs in the cloud through the Watson Machine Learning Bluemix service.
Machine learning methods and algorithms
Supports decision tree, neural networks and regression models. ARMA, ARIMA and exponential smoothing; transfer function with predictors and outlier detection; ensemble and hierarchical models; support vector machine (SVM) and temporal causal modeling (TCM); time series and spatial AR in STP (spatiotemporal prediction). Generative adversarial networks (GANs) and reinforcement learning for deep learning.
Customer Case Studies
See how the Volvo Group used Modeler for predictive maintenance to fulfill uptime commitments.