Day-by-day behavioral datasets have been widely generated in real applications, such as the transaction data in stock market created by accounts day by day. How to effectively mine the informative information from this kind of data is of much challenge for data mining and information retrieval communities. In the day-by-day behavioral dataset, some objects containing the similar iterative behavioral patterns are dramatically different from other individual objects and demonstrate their peculiarities. It is very significant for behavior mining to group these peculiar objects' groups as this has great potential in practice. In this paper, we propose an iterative behavioral pattern-based peculiarity groups mining(IbpPGM) method to group the objects in day-by-day behavioral datasets. In the first stage to tackle the problem, we obtain the iterative behavioral pattern to express the feature of behavior and define the measurement of degree of peculiarity. For mining peculiarity groups, we obtain the feature of an object combined by the features of behavior(i.e.,iterative behavioral patterns),define the measurements of similarity between peculiarity iterative behavioral patterns and group the objects to mine the peculiarity groups based on iterative behavioral patterns. Experiments based on real day-by-day behavioral datasets validate the effectiveness of the proposed approach. |