The most prominent feature of v18.4 is the deepened integration with open-source languages.
Once a model is built, IBM SPSS Modeler 184 offers multiple deployment options:
Modeler 18.4 improves how it connects to big data and cloud storage sources. ibm+spss+modeler+184
Version 18.4 is designed to operate within the IBM Watson Studio ecosystem (on IBM Cloud Pak for Data).
Letās simulate a simple churn prediction project. The most prominent feature of v18
Step 1: Data Source
Drag a Database node. Connect to a SQL Server table containing customer demographics, tenure, monthly charges, and a "Churned" flag.
Step 2: Data Preparation
Step 3: Modeling
Drag an Auto Classifier node. Connect it to the Type node. Run it.
Wait 2ā5 minutes (depending on data size). SPSS Modeler 184 will test:
Step 4: Evaluation
Double-click the Auto Classifier output. Review the Gains Chart and Confusion Matrix. The model with the highest "Overall Accuracy" and "Lift" for the top decile is your champion model. Modeler 18
Step 5: Deployment
Right-click the best model. Select "Save as SQL Script" for SQL Server. This generates a stored procedure that scores new customers in milliseconds.
Time to first insight: Less than 1 hour (with zero code).
Jump to
Every time a new story is published, youāll get an alert straight to your inbox!
Look out for an alert in your inbox the next time a new story is published!
By clicking āSign upā, you agree to receive emails from Business Insider. In addition, you accept Insiderās Terms of Service and Privacy Policy.