Possible bug in the prediction method

[mostafiz67]

I am not sure, I am wrong or what wrong with the package. But something is wrong. For example, when I am trying to calculate the link prediction score for common neighbor it is showing score 1 for node 1 and 8 but the dataset does not have any common neighbor between these 2 nodes. Again, it is not showing any score for node 1 and 6 but we have 1 common neighbor between these 2 nodes.

Code:

df_graph <- read.csv("~/Assignment_3.csv", header = TRUE)
graph <- graph_from_data_frame(df_graph, directed = FALSE)
plot(graph, vertex.label = V(graph)$name)
Common_neighbor <- linkprediction::proxfun(graph, method= "cn", value = "edgelist")

Output

   from to value
1     2  1     2
2     3  1     1
3     4  1     1
4     8  1     1
5     1  2     2
6     3  2     1
7     6  2     1
8     8  2     2
9     1  3     1
10    2  3     1
11    4  3     1
12    7  3     1
13    8  3     1
14    1  4     1
15    3  4     1
16    5  4     2
17    6  4     1
18    7  4     2
19    4  5     2
20    6  5     1
21    7  5     1
22    8  5     1
23    2  6     1
24    4  6     1
25    5  6     1
26    8  6     1
27    3  7     1
28    4  7     2
29    5  7     1
30    1  8     1
31    2  8     2
32    3  8     1
33    5  8     1
34    6  8     1

Reproducible Data:

structure(list(To = c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 4L, 5L, 8L, 
8L), From = c(7L, 3L, 4L, 7L, 4L, 5L, 4L, 6L, 6L, 5L, 6L)), class = "data.frame", row.names = c(NA, 
-11L))

[mbojan]

Thanks @mostafiz67 . Are you using CRAN or the GitHub version of the package?

[mostafiz67]

I am not sure. How can I check?

Most probably CRAN.

linkprediction * 1.0-0 2018-10-19 [1] CRAN (R 3.6.3)