GFD: A Weighted Heterogeneous Graph Embedding Based Approach for Fraud Detection in Mobile Advertising
Algorithm 1
The WMP2vec algorithm.
(1)
Input: The weighted heterogeneous information graph , a meta-path scheme M, walks per node , longest walk length per walk , embedding dimension , neighborhood size
(2)
Output: The latent node embedding
(3)
Initialize , random walk sequence
(4)
fordo
(5)
fordo
(6)
= WeightedMetaPathRandomWalk
(7)
(8)
end
(9)
end
(10)
= HeterogeneousSkipGram
(11)
return
(12)
WeightedMetaPathRandomWalk
(13)
initialize random walk array , weight array ,
(14)
relationship array
(15)
fordo
(16)
fordo
(17)
ifthen
(18)
draw and according to equation (2) with relationship
(19)
(20)
else
(21)
draw and according to equation (3) with relationship
(22)
if does not exist then return
(23)
else
(24)
end
(25)
fordo
(26)
draw and according to equation (3) with relationship
