Examining Socioeconomic and Computational Aspects of Vaccine Pharmacovigilance

Figure 2

Drug name mapping performance: we contrast the number of drugs that were matched in a narrative against its length (sample Pearson correlation coefficient r = 0.69 reflects the effect of noise introduced by longer “dirty” narratives that did not only contain medication names). We notice that the larger the narrative’s length was, the more drugs matched—most matches happened between narratives that were smaller in size.