Not an easy book to read, but for the AI practitioner a source of ideas and important recipes to implement.

This is a challenging book to review. The information covered reminded me of key research principles and their importance as well as highlighted advances in the field. However, the pacing of the book is pedestrian and the repetition throughout, tiresome. When considering the question "Would I recommend it?" I realised I have several to whom I will recommend this book! One is a friend in need of a grounding in research, as she has no experience. Another is a friend who is completing a doctorate and is reviewing research options. It is well compiled, extensive, easy to understand and interesting. If you pick it up intending to read only those aspects relevant to you, you will thoroughly enjoy the experience.

My favourite quotes include:

"More data of the wrong type actually is bad for you. If you are looking for a needle in a haystack it does not help to have a larger haystack. While it is a worthwhile goal to collect as much data as possible, data quality, and knowing which data will address your needs, remain paramount."

"So why focus groups? Aside from detecting disasters, these groups have several valuable applications. You learn about the language that people use in discussing the product, and in particular the terminology that they can understand."

"Experience shows that, for a reasonably sized experiment, 125 per group you want to measure separately is safe and reliable."

Very heavy in Statistics. Artificial Intelligence Marketing and predicting consumer Choice by Steven Struhl has academic tendency. It will make a great book for students and learners.