Abstract: This paper indicates through worldwide analysis that the application of data mining (DM) in petroleum exploration databases is in the initial stage. It briefly describes the structures, functions, algorithms and key techniques of data mining based on the work of pioneer contributors. A case study of well-logging interpretation introduces the processes, methods and application results of DM in concrete terms, proving that the DM used is feasible and practical. In the case study, multiple regression analysis is adopted as a dimension-reduction algorithm, while artificial neural network and support vector machine are employed as mining algorithms for knowledge discovery. Since a great number of petroleum exploration databases (including databanks and datastores) have been or will be built in China, it is time to further develop the techniques of DM that will become powerful tools for petroleum geology research and exploration decision.

Key words: datamining, knowledge discovery, generalized database, dimension-reduction algorithm;mining algorithm, petroleumexploration, fracture prediction