Assessing and Improving Prediction and Classification: Theory and Algorithms in C++

For those seriously interested in prediction and classification Dr. Masters needs no introduction. I have all of his prior books and this one takes the field to the next level. The book deals with topics that are crucial and not to be found in any single book. If you want to be operating at the level of best practices in data mining and machine learning from structured data this book is a must read.I've known Dr. Masters for nearly 20 years and he is an expert of the highest competence, modesty and integrity. If he knows something he will tell you what he knows. But just as freely, he discusses the limits of his knowledge.Explanations are clear and intuitive so even those with modest knowledge of this field will take away much of value from this book.David R. Aronson

May be helpful for the unexperienced reader.

Great insights! Rather than describing different models, it actually focuses on statistical strategies to refine existing models. Must have for people who know how different models work and want to achieve some extra accuracy.

I don't understand how this book was not able to find a publisher. Even I have learned from it (and I'm a professional quant trader), and even where I have not I've seen it positing novel solutions to old problems, or independently reproducing solutions and techniques which are treated as bordering on trade secrets when applied to markets by financial firms.

Do you ever wanted to know more about data mining algorithms? Are you interested in experiences and tricks that are not written in introductory text books? Then, Assessing and Improving Prediction and Classification, by Timothy Masters, is maybe what you need.This advanced book describes topics such as regression and classification method assessment, resampling and combining classifiers. For each algorithm, the books starts with explanations (some equations and graphs when needed) and continues with corresponding C++ code. It is thus not a book to read from its very beginning until its end. The reader will rather pick some preferred chapters to read.Key aspects of data mining are discussed such as carefully selecting the test set. One strength of the book is to explain advanced concepts with very few equations. Masters discusses points that are key to all data mining applications, although often not covered in standard textbooks. On the drawback side, one can mention the low quality of the pictures and graphs in the book. In conclusion, Masters’ book is an excellent support for experienced data miners that prefer plain text descriptions rather than mathematical notations.

Assessing and Improving Prediction and Classification is a very advanced book for predictive analytics. It will not teach you how to create a classifier (predictive model) but how to asses your classifier, how to train different versions of it and how to combine them to achieve even better results.The book cover a lot of ground and make you think a lot. At the end of the book, you have enough ideas on how to increase your predictive analytics to keep you busy a long time.One should note that while the methods are implemented in C++, there is a good documentation of it so it's easy to follow. Something missing in many books containing code excerpt.If you are serious about predictive analytics, if you want to achieve better results, this book is exactly what you need.