“…This argument that has been supported by recent works on privacy-protection (Chow, Golle, & Staddon, 2008;Sánchez, et al, 2013a;Sánchez, Batet, & Viejo, 2013b), which considered the Web as a realistic proxy of social knowledge. In order to compute term probabilities from the Web in an efficient manner, several authors (Sánchez, Batet, Valls, & Gibert, 2010;Turney, 2001) have used the hit count returned by a Web Search Engine (e.g., Bing, Google) when querying the term t. In our approach, term probabilities are computed in this way:…”