SPSS Statistics for Data Analysis and Visualization

Part III Predictive Analytics

Predictive analytics is a big topic, but a timely one. Virtually all of us who are statistically inclined need some techniques borrowed from our Computer Science, Knowledge Discovery, and Machine Learning colleagues in our predictive analytics toolkit. But that doesn’t mean that we need a dedicated tool to do predictive analytics. Dedicated data mining tools like SPSS Modeler are extremely powerful, and they have their place and advantages, but there is much that can be done right in SPSS Statistics. This five-chapter part advances one of the motivating themes of the entire book: What features of SPSS Statistics constitute untapped resources that can potentially transform the way we use SPSS Statistics? Specifically, in this part of the book you learn how SPSS Statistics can be used to do data mining. While not absolutely everyone needs the techniques in Part III, an increasingly large proportion of analysts do. Machine learning is mainstream in a way that it wasn’t even 10 years ago. Familiarity is wise even if one’s day to day statistics work is more traditional.

We don’t always have the benefit of a hypothesis. What are we to do when we have a research question, but we don’t have a hypothesis? For instance, what if we simply want to know who among our customers are most likely to respond to a particular marketing campaign? They are our customers, after all. So we have more information about them than anyone else does. We certainly have some ...