Search

We present MinMax, a new algorithm for mining maximal frequent itemsets(MFI) from a transaction database. It is based on depth-first traversal and iterative. It combines a vertical tidset representation of the database with effective pruning mechanisms. MinMax removes all the non-maximal frequent itemsets to get the exact set of MFI directly, needless to enumerate all the frequent itemsets from smaller ones step by step. It backtracks to the proper ancestor directly, needless level by level . We found MinMax to be more effective than GenMax, a state-of-the-art algorithm for finding maximal frequent item-sets, to prune the search space to get the exact set of MFI.

We propose a new watermarking algorithm for relational database which is based on spread spectrum techniques in this paper. We assign different owners for different identification key to generate the special watermarking signal and insert it into appropriate tuples. At the watermarking detection step, we use the technique of even parity check and majority voting for the watermark accuracy. It is to establish that a false positive judgment is highly unlikely by evaluating the similarity of watermarks between the original and the one gotten from the detection step. Experimental results are also presented to support that it is robust in the present of the various attacks and positive updating such as adding, deleting and modifying.

Previous studies have presented convincing arguments that a frequent sequence mining algorithm should not mine all frequent sequences but only the closed ones because the latter leads to not only more compact yet complete result set but also better efficiency. However, frequent closed sequence mining is still challenging on stand-alone for its large size and high dimension. In this paper, an algorithm, PFCSeq, is presented for mining frequent closed sequence based on distributed-memory parallel machine, in which each processor mines local frequent closed sequence set independently using task parallelism with data parallelism approach, and only two communications are needed except that imbalance is detected. Therefore, time spent in communications is significantly reduced. In order to ensure good load balance among processors, a dynamic workload balance strategy is proposed. Experiments show that it is linearly scalable in terms of database size and the number of processors.