Robustness checking: community detection

[yenchiayi]

• [x] check the robustness of random seeds preliminary guess: there is no much difference. For example, the result of seed 42 is very similar to 52; the top 3 packages within the detected communities are only slightly different in the 17th largest community.

[yenchiayi]

The distribution of Jaccard correlation is described as follows.

Result of 100 random seeds

[yenchiayi]

check the following files:

• visualization_community/comm03_influence_of_random_seeds.csv
• visualization_community/comm03_jaccard_distribution_of_random_seeds.png

[yenchiayi]

Result of 1000 random sees:

[yenchiayi]

I opened a relevant issue regarding the choice of random seeds and invariant output of walktrap algorithm in #25