This week, Eric Siegel discusses the ‘predictive model,’ in Chapter 1 of his book, Predictive Analytics. He defines predictive model in this way:
A mechanism that predicts a behavior of an individual, such as a click, buy, lie, or die. It takes characteristics of the individual as input and provides a predictive score as output. The higher the score, the more likely it is that the individual will exhibit the predicted behavior (34).
The author elaborates on this definition by adding:
Before using a model, we’ve got to build it. Machine learning (computer methods that aid in finding patterns – see Pp. 19-21) the predictive model…Machine learning crunches data to build the model, a brand-new prediction machine. The model is the product of this learning technology – it is itself the very thing that has been learned…Predictive modeling generates the entire model from scratch. All the model’s math, weights, or rules are created automatically by the computer…This automation is the means by which PA builds its predictive power.
From this description we get the picture of a model generated and deployed with the utmost impartiality. However, there is something missing from this rather enthusiastic embrace of the predictive model; somewhere at the beginning an actual person designs the model, often for a company or institution that has an objective or objectives it is trying to accomplish. In other words, models, by necessity, reflect the judgments and priorities of the people creating the model. And an efficient model is not always accurate or insightful
Sample Solution