Graph-Based Node Finding in Big Complex Contextual Social Graphs

Algorithm 1

RN-MCTS.

	Data: CSG ; query graph ; candidate node set ; number of iterations I;
	Perform rapid evaluate to build MCTs of ;
	Use of rapid evaluate to train RNN;
	for iteration in do
	Use probability function to select from ; set the current node ;
	In the upper tree and lower tree of :
	While do
	If is unvisited then
	Mark visited; add the neighbors of in G that into the MCT as the child nodes of ; initial of these child nodes;
	if then
	Break;
	end
	end
	if is a leaf node then
	Break;
	end
	Select the node with the maximum UCT value; update ;
	end
	for each node in do
	Update ; use to train RNN;
	end
	Perform optimization strategy;
	if meet the situation of early terminal then
	Return top-K matches of ;
	end
	end
	Return top-K matches of .