Too much information can be overwhelming, but when it comes to certain types of data that are used to build predictive models to guide decision making there is no such thing as too much data. To determine whether more data is really better for predictive modeling, Enric Junqué de Fortuny and David Martens, University of Antwerp, Belgium, and Foster Provost, New York University, NY, tested nine different applications in which they built models using a particular type of data called fine-grained data, such as observing an individual's behavior in a certain setting. In this article the authors state that "certain telling behaviors may not be observed in sufficient numbers without massive data."
Category: Articles
Click here