From Figure 5, we see the lowest average relevance scores in all the cases where the Low Quality
Parameters (see section 4.3) were used to construct the feature vectors. Further, we observed that the
highest scores were achieved for all the cases where the Medium Quality Parameters were used. Figure 6
depicts the relevance scores for the entities used with the MQP and the HQP parameters relative to the entities where the LQP were used in percentage. We observed the same pattern; in cases where the FVs
were created using the MQP the score was highest.