CDP Machine Learning Models: What You Need To Know

Few business people today would question the value of machine learning. Its unique benefit is the ability to build models by training itself, rather than relying on help from expert statisticians. This lets it create reliable models more quickly and at lower cost than methods that depend on human expertise. The result is that many more models can be built and existing models can more easily be updated as conditions change.

This enables a vast array of applications to take advantage of data-based predictions. The specifics depend on the particular business. Typical examples include:

• Product category affinity models for retailers, enabling personalization systems to offer each customer products in the category they are most likely to purchase
• Mortgage assessment models for banks, enabling financial institutions to offer each customer the lending terms that best balance risk and profitability
• Rehospitalization predictions for health care providers, enabling organizations to focus compliance support on patients who are most likely to need help
• Customer segment assignments for telecom providers, enabling them to offer the most appropriate service plans to customers with different needs
• Response prediction models for hospitality companies, enabling them to minimize the number of discount offers needed to sell unused inventory

White Paper from