COMMunication
RESearch.NET

User Tools

Site Tools

You are here: index » krackhardt_datasets
Trace:
krackhardt_datasets

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
krackhardt_datasets [2019/12/04 08:37] – [correlation matrix out of the combined matrix (friend and advice)] hkimscilkrackhardt_datasets [2019/12/13 14:11] (current) – hkimscil
Line 65: Line 65:
  
 <code> <code>
-krack_friend <- delete.edges(krack_full, E(krack_full)[E(krack_full)$friendship_tie==0])+krack_friend <- delete_edges(krack_full, E(krack_full)[E(krack_full)$friendship_tie==0])
 summary(krack_friend) summary(krack_friend)
 krack_friend[] krack_friend[]
  
-krack_advice <- delete.edges(krack_full, E(krack_full)[E(krack_full)$advice_tie==0])+krack_advice <- delete_edges(krack_full, E(krack_full)[E(krack_full)$advice_tie==0])
 summary(krack_advice) summary(krack_advice)
 krack_advice[] krack_advice[]
    
-krack_reports_to <- delete.edges(krack_full, E(krack_full)[E(krack_full)$reports_to_tie==0])+krack_reports_to <- delete_edges(krack_full, E(krack_full)[E(krack_full)$reports_to_tie==0])
 summary(krack_reports_to) summary(krack_reports_to)
 krack_reports_to[] krack_reports_to[]
Line 213: Line 213:
  
 <code> <code>
-# Next, we'll use the same procedure to add social-interaction+# Next, we'll use the same procedure to add advice
 # information. # information.
 krack_advice_matrix_row_to_col <- get.adjacency(krack_advice, attr='advice_tie') krack_advice_matrix_row_to_col <- get.adjacency(krack_advice, attr='advice_tie')
Line 226: Line 226:
 krack_advice_matrix <- rbind(krack_advice_matrix_row_to_col, krack_advice_matrix_col_to_row) krack_advice_matrix <- rbind(krack_advice_matrix_row_to_col, krack_advice_matrix_col_to_row)
 krack_advice_matrix krack_advice_matrix
- +</code> 
 + 
 + 
 +<code> 
 +krack_friend_matrix_row_to_col <- get.adjacency(krack_friend, attr='friendship_tie') 
 +krack_friend_matrix_row_to_col 
 + 
 +# To operate on a binary graph, simply leave off the "attr"  
 +# parameter: 
 +krack_friend_matrix_row_to_col_bin <- get.adjacency(krack_friend) 
 +krack_friend_matrix_row_to_col_bin 
 + 
 +# For this lab, we'll use the valued graph. The next step is to  
 +# concatenate it with its transpose in order to capture both  
 +# incoming and outgoing task interactions. 
 +krack_friend_matrix_col_to_row <- t(as.matrix(krack_friend_matrix_row_to_col)) 
 +krack_friend_matrix_col_to_row 
 + 
 +krack_friend_matrix <- rbind(krack_friend_matrix_row_to_col, krack_friend_matrix_col_to_row) 
 +krack_friend_matrix 
 +</code> 
 + 
 + 
 +<code>  
 +# ra (ar)
 krack_reports_to_advice_matrix <- rbind(krack_reports_to_matrix, krack_advice_matrix) krack_reports_to_advice_matrix <- rbind(krack_reports_to_matrix, krack_advice_matrix)
 krack_reports_to_advice_matrix krack_reports_to_advice_matrix
 +
 +# fa
 +krack_friend_advice_matrix <- rbind(krack_friend_matrix, krack_advice_matrix)
 +krack_friend_advice_matrix
 +
 +# fr
 +krack_friend_reports_to_matrix <- rbind(krack_friend_matrix, krack_reprots_to_matrix)
 +krack_friend_reports_to_matrix
 +
 +
 +# far
 +krack_friend_advice_reports_to_matrix <- rbind(krack_friend_advice_matrix, krack_reports_to_matrix)
 +krack_friend_advice_reports_to_matrix
 </code> </code>
 +
  
 <code> <code>
Line 238: Line 276:
 krack_reports_to_advice_cors <- cor(as.matrix(krack_reports_to_advice_matrix)) krack_reports_to_advice_cors <- cor(as.matrix(krack_reports_to_advice_matrix))
 krack_reports_to_advice_cors krack_reports_to_advice_cors
 +
 +krack_friend_advice_cors <- cor(as.matrix(krack_friend_advice_matrix))
 +krack_friend_advice_cors
 +
 +krack_friend_reports_to_cors <- cor(as.matrix(krack_friend_reports_to_matrix))
 +krack_friend_reports_to_cors
 +
 +krack_friend_advice_reports_to_cors <- cor(as.matrix(krack_friend_advice_reports_to_matrix))
 +krack_friend_advice_reports_to_cors
 +
 +
 </code> </code>
  
Line 246: Line 295:
 # or equal to 0; thus, highly dissimilar (i.e., negatively  # or equal to 0; thus, highly dissimilar (i.e., negatively 
 # correlated) actors have higher values. # correlated) actors have higher values.
-dissimilarity <- 1 - krack_reports_to_advice_cors +dissimilarity_ra <- 1 - krack_reports_to_advice_cors 
-krack_reports_to_dist <- as.dist(dissimilarity) +krack_reports_to_advice_dist <- as.dist(dissimilarity_ra) 
-krack_reports_to_dist+krack_reports_to_advice_dist
    
 +dissimilarity_fa <- 1 - krack_friend_advice_cors
 +krack_friend_advice_dist <- as.dist(dissimilarity_fa)
 +krack_friend_advice_dist
 +
 +dissimilarity_rf <- 1 - krack_reports_to_friend_cors
 +krack_reports_to_friend_dist <- as.dist(dissimilarity_rf)
 +krack_reports_to_friend_dist
 +
 +dissimilarity_far <- 1 - krack_friend_advice_reports_to_cors
 +krack_friend_advice_reports_to_dist <- as.dist(dissimilarity_far)
 +krack_friend_advice_reports_to_dist
 +
 +
 +
 # Note that it is also possible to use dist() directly on the  # Note that it is also possible to use dist() directly on the 
 # matrix. However, since cor() looks at associations between  # matrix. However, since cor() looks at associations between 
Line 1374: Line 1437:
 # different numbers of clusters and assign vertices to clusters # different numbers of clusters and assign vertices to clusters
 # as appropriate. For example: # as appropriate. For example:
-cutree(krack_reports_to_advice_hclust, k=2) +cutree(krack_friend_advice_hclust, k=2) 
- +</code> 
 + 
 +<code> 
 +> cutree(krack_friend_advice_hclust, k=2) 
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21  
 + 1  2  1  1  1  2  1  1  1  1  2  2  1  2  1  1  2  1  1  1  1  
 +>  
 +>  
 +</code> 
 + 
 +<code> 
 # Now we'll try to figure out the number of clusters that best  # Now we'll try to figure out the number of clusters that best 
 # describes the underlying data. To do this, we'll loop through # describes the underlying data. To do this, we'll loop through
Line 1399: Line 1472:
 # set a variable for our number of vertices. # set a variable for our number of vertices.
 clustered_observed_cors = vector() clustered_observed_cors = vector()
-num_vertices = length(V(krack_reports_to)) +num_vertices = length(V(krack_advice)) 
- + 
 # Next, we loop through the different possible cluster  # Next, we loop through the different possible cluster 
 # configurations, produce matrices of within- and between- # configurations, produce matrices of within- and between-
Line 1407: Line 1481:
    
 # pdf("6.3_m182_studentnet_task_social_clustered_observed_corrs.pdf") # pdf("6.3_m182_studentnet_task_social_clustered_observed_corrs.pdf")
-clustered_observed_cors <-clustConfigurations(num_vertices, krack_reports_to_advice_hclust, krack_reports_to_advice_cors)+clustered_observed_cors <-clustConfigurations(num_vertices, krack_friend_advice_hclust, krack_friend_advice_cors)
 clustered_observed_cors clustered_observed_cors
 plot(clustered_observed_cors$correlations) plot(clustered_observed_cors$correlations)
 # dev.off() # dev.off()
- +
 clustered_observed_cors$correlations clustered_observed_cors$correlations
 +</code>
 +
 +<code>
 +> clustered_observed_cors <-clustConfigurations(num_vertices, krack_friend_advice_hclust, krack_friend_advice_cors)
 +Warning message:
 +In cor(as.vector(d[g1[i], , ]), as.vector(d[g2[j], , ]), use = "complete.obs") :
 +  표준편차가 0입니다
 +> clustered_observed_cors
 +$label
 +[1] "number of clusters:  1"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1 
 +
 +$correlation
 +[1] NA
 +
 +$label
 +[1] "number of clusters:  2"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  1  1  1  2  1  1  1  1  2  2  1  2  1  1  2  1  1  1  1 
 +
 +$correlation
 +[1] 0.4896211
 +
 +$label
 +[1] "number of clusters:  3"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  2  1  1  3  3  2  2  3  2  3  1  2  3  3  1  1 
 +
 +$correlation
 +[1] 0.5944114
 +
 +$label
 +[1] "number of clusters:  4"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  2  1  1  3  4  2  2  3  2  3  1  2  4  3  1  1 
 +
 +$correlation
 +[1] 0.6398013
 +
 +$label
 +[1] "number of clusters:  5"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  2  4  1  3  5  2  2  3  2  3  1  2  5  3  4  4 
 +
 +$correlation
 +[1] 0.6538231
 +
 +$label
 +[1] "number of clusters:  6"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  4  5  1  3  6  4  2  3  4  3  1  4  6  3  5  5 
 +
 +$correlation
 +[1] 0.6723019
 +
 +$label
 +[1] "number of clusters:  7"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  4  5  1  3  6  4  2  3  4  3  1  4  6  3  5  7 
 +
 +$correlation
 +[1] 0.7019599
 +
 +$label
 +[1] "number of clusters:  8"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  4  5  1  3  6  4  2  3  4  3  1  7  6  3  5  8 
 +
 +$correlation
 +[1] 0.727137
 +
 +$label
 +[1] "number of clusters:  9"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  1  3  4  5  1  3  6  4  2  7  4  7  1  8  6  7  5  9 
 +
 +$correlation
 +[1] 0.7743714
 +
 +$label
 +[1] "number of clusters:  10"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  4  3  7  5  2  8  5  8  4  9  7  8  6 10 
 +
 +$correlation
 +[1] 0.7919439
 +
 +$label
 +[1] "number of clusters:  11"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  4  3  7  5  2  8  5  8  4  9 10  8  6 11 
 +
 +$correlation
 +[1] 0.8093965
 +
 +$label
 +[1] "number of clusters:  12"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  5  2  9  5  9  7 10 11  9  6 12 
 +
 +$correlation
 +[1] 0.8445199
 +
 +$label
 +[1] "number of clusters:  13"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  5  2  9  5 10  7 11 12  9  6 13 
 +
 +$correlation
 +[1] 0.8700886
 +
 +$label
 +[1] "number of clusters:  14"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  9  2 10  5 11  7 12 13 10  6 14 
 +
 +$correlation
 +[1] 0.8844067
 +
 +$label
 +[1] "number of clusters:  15"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  9  2 10 11 12  7 13 14 10  6 15 
 +
 +$correlation
 +[1] 0.9115517
 +
 +$label
 +[1] "number of clusters:  16"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  9  2 10 11 12  7 13 14 10 15 16 
 +
 +$correlation
 +[1] 0.9403353
 +
 +$label
 +[1] "number of clusters:  17"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  9  2 10 11 12  7 13 14 15 16 17 
 +
 +$correlation
 +[1] 0.9502702
 +
 +$label
 +[1] "number of clusters:  18"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  3  8  9 10 11 12 13  7 14 15 16 17 18 
 +
 +$correlation
 +[1] 0.9633198
 +
 +$label
 +[1] "number of clusters:  19"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  8  9 10 11 12 13 14  7 15 16 17 18 19 
 +
 +$correlation
 +[1] 0.9762881
 +
 +$label
 +[1] "number of clusters:  20"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  3  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 
 +
 +$correlation
 +[1] 0.9895545
 +
 +$label
 +[1] "number of clusters:  21"
 +
 +$clusters
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 + 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 
 +
 +$correlation
 +[1] 1
 +
 +$correlations
 + [1]        NA 0.4896211 0.5944114 0.6398013 0.6538231 0.6723019 0.7019599 0.7271370 0.7743714 0.7919439 0.8093965 0.8445199 0.8700886 0.8844067 0.9115517 0.9403353
 +[17] 0.9502702 0.9633198 0.9762881 0.9895545 1.0000000
 +
 +> plot(clustered_observed_cors$correlations)
 +> 
 +> clustered_observed_cors$correlations
 + [1]        NA 0.4896211 0.5944114 0.6398013 0.6538231 0.6723019 0.7019599 0.7271370 0.7743714 0.7919439 0.8093965 0.8445199 0.8700886 0.8844067 0.9115517 0.9403353
 +[17] 0.9502702 0.9633198 0.9762881 0.9895545 1.0000000
 +>
 +</code>
 +{{:pasted:20191204-081317.png}}
 +
 +===== decision on # of clusters  =====
 +
 +<code> 
 # From a visual inspection of the correlation matrix, we can  # From a visual inspection of the correlation matrix, we can 
 # decide on the proper number of clusters in this network.  # decide on the proper number of clusters in this network. 
-# For this network, we'll use 4. (Note that the 1-cluster +# For this network, we'll use 6. (Note that the 1-cluster 
 # solution doesn't appear on the plot because its correlation  # solution doesn't appear on the plot because its correlation 
 # with the observed correlation matrix is undefined.) # with the observed correlation matrix is undefined.)
 num_clusters = 4 num_clusters = 4
-clusters <- cutree(krack_reports_to_advice_hclust, k = num_clusters)+clusters <- cutree(krack_friend_advice_hclust, k = num_clusters)
 clusters clusters
-  + 
-cluster_cor_mat <- clusterCorr(krack_reports_to_advice_cors, +cluster_cor_mat <- clusterCorr(krack_friend_advice_cors, clusters) 
-                                            clusters) +round(cluster_cor_mat,2) 
-cluster_cor_mat +
- +
 # Let's look at the correlation between this cluster configuration  # Let's look at the correlation between this cluster configuration 
 # and the observed correlation matrix. This should match the  # and the observed correlation matrix. This should match the 
 # corresponding value from clustered_observed_cors above. # corresponding value from clustered_observed_cors above.
-gcor(cluster_cor_mat, krack_reports_to_advice_cors) +gcor(cluster_cor_mat, krack_friend_advice_cors) 
- + 
 +</code> 
 + 
 +<code> 
 ##################### #####################
 # Questions: # Questions:
Line 1437: Line 1747:
 ##################### #####################
      
- 
- 
 ### NOTE ON DEDUCTIVE CLUSTERING ### NOTE ON DEDUCTIVE CLUSTERING
  
Line 1455: Line 1763:
 # You could then examine the fitness of this cluster configuration # You could then examine the fitness of this cluster configuration
 # as follows: # as follows:
-deductive_cluster_cor_mat <- generate_cluster_cor_mat( +deductive_cluster_cor_mat <- generate_cluster_cor_mat(krack_friend_advice_cors, deductive_clusters)
-  krack_reports_to_advice_cors, +
-  deductive_clusters)+
 deductive_cluster_cor_mat deductive_cluster_cor_mat
-gcor(deductive_cluster_cor_mat, krack_reports_to_advice_cors)+gcor(deductive_cluster_cor_mat, krack_friend_advice_cors)
  
 ### END NOTE ON DEDUCTIVE CLUSTERING ### END NOTE ON DEDUCTIVE CLUSTERING
Line 1467: Line 1773:
 # networks. # networks.
  
-# Task valued +# Friendship valued 
-task_mean <- mean(as.matrix(krack_reports_to_matrix_row_to_col)_ +friend_mean <- mean(as.matrix(krack_friend_matrix_row_to_col)) 
-task_mean+friend_mean
  
-task_valued_blockmodel <- blockmodel(krack_reports_to_matrix_row_to_col, clusters) +friend_valued_blockmodel <- blockmodel(krack_friend_matrix_row_to_col, clusters) 
-task_valued_blockmodel+friend_valued_blockmodel
  
-# Task binary +# friend binary 
-task_density <- graph.density(krack_reports_to) +friend_density <- graph.density(krack_friend) 
-task_density+friend_density
  
-task_binary_blockmodel <- blockmodel(as.matrix(krack_reports_to_matrix_row_to_col_bin), clusters) +friend_binary_blockmodel <- blockmodel(as.matrix(krack_friend_matrix_row_to_col_bin), clusters) 
-task_binary_blockmodel+friend_binary_blockmodel
  
  
-# Social valued+# advice valued
 advice_mean <- mean(as.matrix(krack_advice_matrix_row_to_col)) advice_mean <- mean(as.matrix(krack_advice_matrix_row_to_col))
 advice_mean advice_mean
Line 1489: Line 1795:
 advice_valued_blockmodel advice_valued_blockmodel
  
-# Social binary+# advice binary
 advice_density <- graph.density(krack_advice) advice_density <- graph.density(krack_advice)
 advice_density advice_density
Line 1495: Line 1801:
 advice_binary_blockmodel <- blockmodel(as.matrix(krack_advice_matrix_row_to_col_bin), clusters) advice_binary_blockmodel <- blockmodel(as.matrix(krack_advice_matrix_row_to_col_bin), clusters)
 advice_binary_blockmodel advice_binary_blockmodel
 +
 +
  
 # We can also permute the network to examine the within- and  # We can also permute the network to examine the within- and 
Line 1501: Line 1809:
 cluster_cor_mat_per <- permute_matrix(clusters, cluster_cor_mat) cluster_cor_mat_per <- permute_matrix(clusters, cluster_cor_mat)
 cluster_cor_mat_per cluster_cor_mat_per
 +</code>
  
 +<code>
 ##################### #####################
 # Questions: # Questions:
krackhardt_datasets.1575416255.txt.gz · Last modified: 2019/12/04 08:37 by hkimscil
Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki