GAMLSS for Big Data: Roc Curve prediction using Twitter data

During last years, Big Data appears as one of the most innovative and growing scientific area of interest. In this field, finding reliable methods to make accurate predictions represents one of the most inspirational challenges. The way to make prediction in the following paper is the use of ROC (Receiver Operating Characteristic) Curve, a binary prediction tool, often used for medical tests. The attention is focused in particular on the implementation of ROC Curve in GAMLSS (Generalized Additive Models for Location Scale and Shape), semi-parametric models suitable for huge and flexible dataset. An application will be shown where the class of GAMLSS is applied to Twitter data in order to predict number of interactions for a tweet given a set of explanatory variables.