Ranking the importance of node in a complex network
- g++>=4.8
- python==2.7
- chainer>=5.1.0
- numpy>=1.15.4
- matplotlib>=2.2.3
We tried to optimize the PageRank algorithm using learning to rank.
We use LETOR4.0 dataset to calculate PageRank values, you can download the Gov2Link.rar from here.
To get PageRank values from 25,000,000 nodes, please make sure that IvtLinks. txt and PageRank.cpp are in the same directory.
g++ ./code/PageRank/PageRank.cpp
./a.out
To get dataset without PageRank value
python ./code/RankInNet/bin/build/remove_pr_value.py
To get dataset with PageRank value
python ./code/RankInNet/bin/build/add_pr_value.py
Train without PageRank value
python ./code/RankInNet/bin/train_without_pr.py
Train with PageRank value
python ./code/RankInNet/bin/train.py