*4.4. Cluster Detection Indenpent Analysis*

In a second analysis, the relationships between the publications analysed will be detected. This analysis is independent of the ASJC's Scopus classification done in previous section. In this case the analysis was done with a cluster detection algorithm that contains the software Gephi. Thus, the clusters have been obtained according to the relationships that exist between the publications. Figure 12 shows a color-coded according to the twenty-two clusters cluster obtained. The weight of the cluster reflects in ratio the significance of this set of publications in the whole network of relations. Once the clusters are established, all the keywords are extracted from all the publications in that cluster. Then, the frequency of each keyword that is found in each cluster is calculated as an index of its importance within that cluster. Tables 6–11 show a list of the main keywords for the leading clusters found, up to 5% of weight. The proposed name for each cluster was made according to the keywords of this cluster.



**Figure 12.** Network of the relationship between publications that are cited in patents in the field of medicine according to the subcategories of medicine of the ASJC.



**Table 8.** DNA Repair. Cluster (11), weigh 8.91%.



**Table 9.** Human leukocyte antigen (HLA). Cluster (5), weigh 7.74%.

**Table 10.** Alzheimer Disease. Cluster (19), weigh 5.84%.


**Table 11.** Carcinoma. Cluster (6), weigh 5.15%.


The advantage of this second analysis is that it allows to detect which specific medical topics are being transferred to patents. Thus, the leading topics obtained were: neoplasms, leukemia, DNA repair, human leukocyte antigen, Alzheimer disease, and carcinoma.
