Home

Nowcasting is used for the prediction of the present, the near future and the near past in economy. It is used to assess the state of an economy (GPD for example). Nowcasting exploits a large quantity of data to determine the tendencies of an economic indicator.

This framework has been applied on websites and particularly on prediction of human behavior. For example, there are robots which can play poker and chess by anticipating the future movements of its opponent. These robots learn by observing the game and the movements of its opponent, it is machine learning. But what interests us is the algorithm of prediction.

Prediction of human behavior intervene in criminality (percentage of risk of recidive of a person), in consumption… It is the consumption dimension we will approach here.

Data are collected but users don’t know how, by who and which data. Moreover, there are private data collected without the user’s agreement. However, these data are needed to the prediction and this prediction is useful to users. The question is if it is possible to predict consumer behavior while protecting users’ data.