The company we keep. Using hemodialysis social network data to classify patients’ kidney transplant attitudes with machine learning algorithms

Table 3 Network statistics and attitude towards kidney transplantation

Network Variable	Positive Attitude N = 66 Mean (SD)	Negative Attitude N = 44 Mean (SD)	p value
Degree centrality	0.019 (0.019)	0.014 (0.014)	0.13
Eigenvector Centrality	0.052 (0.102)	0.021 (0.051)	0.08
Closeness Centrality	0.048 (0.040)	0.037 (0.033)	0.14
Betweenness Centrality	0.007 (0.012)	0.003 (0.006)	0.02
Clustering Coefficient	0.148 (0.294)	0.155 (0.308)	0.91
Triangles	0.606 (1.179)	0.523 (1.055)	0.73

Table 3 shows the association between network statistics and attitude towards kidney transplantation. The statistics have been normalized to the statistics of clinic’s network thus the mean value is zero. Degree centrality is the number of relationships a network member has. Eigenvector centrality is how many relationships a network member has to other network member with lots of relationships. Closeness centrality is a measurement of a participant’s distance by relationships to other network members. Betweenness centrality is a measure of how many unique paths between network members must pass through. Clustering Coefficient is the proportion of actual relationships versus possible relationships among a person’s direct network. Triangles is the number of mutual relationships a network member shares with their other network members. The p value is calculated via randomization test with 10,000 permutations [27]. Standard Deviation (SD)

ISSN: 1471-2369