This book will give you the critical information you need to create, use, and validate simple predictive models, and it will suggest the types of real-world business problems you can solve with those models.
It is designed to be as simple as possible, providing basic, practical, and immediately applicable information for business users new to the world of predictive modeling.
In summary:
An introduction to and some fundamentals for good analysis
A process outline to make analysis quick and effective
A description of some of the most used predictive models and methods, and how they relate to business questions
Comprehensive "How To" sections, including step-by-step Excel tutorials and common pitfalls to avoid
Our approach is as follows:
First, introduce analysis fundamentals. These are the basics of doing good and accurate analysis, and it will be important to keep these principles in mind as you create predictive models.
Second, explain the process that will allow you to follow some easy, predefined steps to creating your own predictive models. This is a "big-picture" process flow meant to give you a basic procedure to follow no matter what type of predictive model you need to create.
Last, this guide gives you an in-depth look into various predictive modeling techniques, organized according to the type of data you have and the type of questions you're trying to answer. This section makes up the bulk of the book, and the explanation of each model tells you what the predictive model looks like, what it can be used for, the assumptions necessary to use the model, a process to follow to create it (including step-by-step instructions in Excel), an explanation of some common errors to watch for, and a section on analyzing your results.
The modeling process you will learn is as follows:
1. Choose a predictive model according to the business question.
2. Check to see if all the conditions for the model are met.
3. Carry out the analysis.
4. Check for statistical significance and fit.
5. Validate the predictive model.
6. Refine the predictive model.
The basic models we go over in this text:
General Regression (linear, multivariate, exponential, logarithmic, polynomial, time series)
Logistic Regression
ANOVA (t-test, one and two-way ANOVA)
Chi-Square
These models cover four common prediction cases you will encounter:
Predict a numerical outcome with numerical explanatory variables
Predict a yes or no outcome with numerical explanatory variables
Predict a numerical outcome with categorical explanatory variables
Predict a categorical outcome with categorical explanatory variables
What you will not get in this book:
Complex statistical explanations
Complex math
Complex predictive models (read: machine learning is not covered)
Python, R, or other coding languages used for modeling
What you will get in this book:
Also, this book may or may not mention wombats.
