There are at least four kinds of books within data mining field. The first category focuses on theory and algorithms. The second one deals with specific tools and languages. The third class is for Management and C-level. The fourth group is concerned with practical applications and guidelines. Dean Abbott’s book, Applied Predictive Analytics – Principles and Techniques for the Professional Data Analyst, fall in the latter category. This is good news since there are very few books available of this type and Dean’s book achieves its objective: giving data miners practical advices to solve their challenges, using the CRISP-DM methodology.
Applied Predictive Analytics describes all CRISP-DM steps. It also discusses descriptive/predictive analytics, association rules, ensemble learning and text mining. I was just surprised by the absence of Support Vector Machines (SVM), which is a powerful technique often used in industry. For each CRISP-DM step, the author describes most possible issues and related ways to solve them.
To be noted the interesting discussion about related fields such as statistics and Business Intelligence. I also particularly liked end of chapters which are giving real and practical advices for specific methods. In conclusion, Dean’s book is written for medium to expert analytics professionals. It will teach advices and highlight pitfalls to avoid. Applied Predictive Analytics is a comprehensive review of the field with a practical perspective, which is rare enough to be noted.