TIBCO StreamBase® provides native TIBCO® Enterprise Runtime for R integration, allowing you perform all of the following in real-time.
- Run R Scripts, including scripts utilizing functions from CRAN libraries
- Load, Generate, and Export R Models, including models based on CRAN libraries
- Score new Data and new Transactions in TIBCO Enterprise Runtime for R
- Export Model Parameters from R models to be scored directly in StreamBase
- Watch for models exported from other systems (e.g. RStudio, R based TIBCO Spotfire® Models) and load these models. Models can be updated in real-time.
- Evaluting Models in real-time to see when Model accuracy has degraded, then triggering a model retraining process.
StreamBase natively pre and post processes the data, performing operations such as data cleaning, transformations, and aggregations before the TIBCO Enterprise Runtime for R script/scoring, and then processing the result form the transaction being scored against the predictive model. Post processing includes using output of the scoring process to create alerts, notifications, trigger business processes, alert windowing, writing to a range of relational and non relational data sources, and other actions based on the model scoring process. StreamBase provides multiple alerting and notification metrics, including control charting, rules, probabilities from models, and others.
Learn more about Streambase