Transaction data in domains such as trading records from online financial or other markets, logistics delivery registers, and many others, are being accumulated at an increasing rate. In this type of data each transaction has a complex format, being usually associated with attributes such as time, numerical quantity, parties involved, and so on. Performing data mining on trace record data of complex transactions may enable the extraction of knowledge about implicit relationships which will benefit the community in different ways, for example by improving market efficiency and oversight, or detecting scheduling bottlenecks. However, the size of data sets of this type is usually enormous, and therefore in order to perform searching or mining techniques considerations of efficiency are often more important than correctness. In this paper we develop a framework to embed different methods to speed up search algorithms with the goal of detecting cycles in trace record data that fit a given constraint predicate on the amount of transaction quantities that can flow in a direction. The method is shown to improve significantly on a naive approach, and suggests a number of directions for further work.
[1]
James C. Tiernan,et al.
An efficient search algorithm to find the elementary circuits of a graph
,
1970,
CACM.
[2]
Albert,et al.
Emergence of scaling in random networks
,
1999,
Science.
[3]
Hannu Toivonen,et al.
Discovery of frequent DATALOG patterns
,
1999,
Data Mining and Knowledge Discovery.
[4]
Lawrence B. Holder,et al.
Substructure Discovery Using Minimum Description Length and Background Knowledge
,
1993,
J. Artif. Intell. Res..
[5]
Takeo Yamada,et al.
Finding all the negative cycles in a directed graph
,
2002,
Discret. Appl. Math..
[6]
Hiroshi Motoda,et al.
Graph-based induction as a unified learning framework
,
1994,
Applied Intelligence.
[7]
Girish Keshav Palshikar,et al.
Collusion set detection using graph clustering
,
2008,
Data Mining and Knowledge Discovery.
[8]
Takashi Washio,et al.
State of the art of graph-based data mining
,
2003,
SKDD.
[9]
Christos Faloutsos,et al.
Graph mining: Laws, generators, and algorithms
,
2006,
CSUR.