Towards bridging theory and practice

Abstract

Graph is a ubiquitous structure representing entities and their relationships applied in many areas such as social networks, web graphs, and biological networks. One of the fundamental tasks in graph analytics is to investigate the relations between two vertices (e.g., users, items and entities) such as how a vertex A influences another vertex B , or to what extent A and B are similar to each other, based on the graph topology structure. For this purpose, we study the problem of hop-constrained s-t simple path enumeration in this paper, which aims to list all simple paths from a source vertex s to a target vertex t with hop-constraint k. We first propose a polynomial delay algorithm, namely BC-DFS, based on barrier-based pruning technique. Then a join-oriented algorithm, namely JOIN, is designed to further enhance the query response time. On the theoretical side, BC-DFS is a polynomial delay algorithm with O ( km ) time per output where m is the number of edges in the graph. This time complexity is the same as the best known theoretical result for the polynomial delay algorithms of this problem. On the practical side, our comprehensive experiments on 15 real-life networks demonstrate the superior performance of the BC-DFS algorithm compared to the state-of-the-art techniques. It is also reported that the JOIN algorithm can further significantly enhance the query response time.

Parallel path detection for fraudulent accounts in banks based on graph analysis

“…In Table 3 , we compared the time efficiency between DFS and our algorithm. DFS ( Grossi, Marino & Versari, 2018 ; Peng et al, 2019 ; Rizzi, Sacomoto & Sagot, 2014 ) is a brute-force search algorithm that finds all nodes that may be passed from the start node to the target node by depth-first search. There are two approaches to solve the DFS problem: recursive and non-recursive.…”

“…To introduce our algorithm, we first review the conventional serial algorithm that uses depth-first search to enumerate all nodes between and . The sequential DFS algorithm explores each branch of the graph as far as possible from the start node until it reaches a dead end, then backtracks to another branch ( Grossi, Marino & Versari, 2018 ; Peng et al, 2019 ; Rizzi, Sacomoto & Sagot, 2014 ). When one of the neighbours of the current node can reach the target node, all nodes within the same SCC are marked as pathnodes.…”