Data Mining Book Review: Predictive Analytics
I had the pleasure to shortly meet Eric Siegel at a Predictive Analytics World conference in 2011. In addition to being the founder of PAW, Eric is also the author of the delightful book Predictive Analytics: the power to predict who will click, buy, lie or die. As written by Thomas Davenport in the foreword, this book is “a good counterpoint to the work of Nassim Nicholas Taleb”. For example, Fooled by Randomness is excellent, but pessimistic. On the opposite, Eric’s book is optimistic: it shows dozens of successful predictive analytics applications.
The public of the book is rather beginner or experts with a need to get ideas about possible predictive analytics applications. As expected from the title, the topic of descriptive analytics (outlier, clustering, etc.) is not covered. However, there is way enough examples of predictive analytics to fill a book. One of the main quality of the book is to cover a (very) wide range of predictive analytics examples: marketing, health care, fraud, finance, human resources, etc.
Eric’s book really makes analytics accessible. It also goes deeper with surprising analytics insights. To be noted an interesting discussion about false positive in crime prediction. In addition to covering a subject such as correlation/causation mix, it lists several examples of it. The only point I may disagree with the author is about the very definition of predictive analytics. Eric defines it using the notion of individual’s behavior. To me, the word by word definition is: ANALYsis using statisTICS (ANALYTICS) for predictive purposes. Topics such as weather forecasting, predictive maintenance, etc. are examples of predictive analytics (according to the textual definition). To a larger extend, any classification/regression task in data mining and machine learning is part of predictive analytics.
To conclude, Eric’s book is a journey in the world of predictive analytics. The book is delightful from the first to the last word. For example, the Watson Jeopardy story is very well explained. Beginners: jump into the field of predictive analytics with Eric’s book! Experts, get a fresh view of the field and gather ideas for your own applications of predictive analytics!