A Project-based PMiner Algorithm in Uncertain Databases

Existing works considered the level-wise method to find the high average-utility sequential patterns at different levels, which may cause higher computations and the huge memory requirements. To solve the previous problem of high average-utility sequential pattern mining, a project-based algorithm called PMiner is presented here to speed up the mining efficiency compared to the previous work. Also, the developed PMiner concerns not only the average-utility on the sequential data but also the data in the uncertain environment (i.e., Internet of Things). Several experiments are then evaluated to show that the proposed algorithm achieves promising performance compared to the level-wise model.