Logistical Regression Model
With steps 1-4 with HDP, we were able to build a logistical regression model. The model was then exported into PMML. The below diagram illustrates the features, coefficients and output of the model.
Getting Started with Streaming Analytics
Also available as:
- 1. Building an End-to-End Stream Application
- 2. Prepare Your Environment
- 3. Creating a Dataflow Application
- 4. Creating a Stream Analytics Application
- Create a Service Pool and Environment
- Create Your First Application
- Creating and Configuring the Kafka Source Stream
- Connecting Components
- Joining Multiple Streams
- Filtering Events in a Stream using Rules
- Using Aggregate Functions over Windows
- Implementing Business Rules on the Stream
- Transforming Data using a Projection Processor
- Creating Alerts with Notifications Sink
- Streaming Alerts to an Analytics Engine for Dashboarding
- Streaming Violation Events to an Analytics Engine for Descriptive Analytics
- Streaming Violation Events into a Data Lake and Operational Data Store
- 5. Deploy an Application
- 6. Stream Operations
- 7. Advanced: Doing Predictive Analytics on the Stream
- Logistical Regression Model
- Export the Model into SAM's Model Registry
- Enrichment and Normalization of Model Features
- Setting up your Enrichment Store and Building Custom UDFs and Processors
- Upload Custom Processors and UDFs for Enrichment and Normalization
- Scoring the Model in the Stream using a Streaming Split Join Pattern
- Streaming Split Join Pattern
- Score the Model using the PMML Processor and Alert
- 8. Creating Visualizations Using Superset