Search

Hiding high utility and frequent itemset is the method used to preserve sensitive knowledge from being revealed by pattern mining process. Its goal is to remove sensitive high utility and frequent itemsets from a database before sharing it for data mining purposes while minimizing the side effects. The current methods succeed in the hiding goal but they cause high side effects. This paper proposes a novel algorithm, named HSUFIBL, that applies a heuristic for finding victim item based on the constrained intersection lattice theory. This algorithm specifies exactly the condition that allows the application of utility reduction or support reduction method, the victim item, and the victim transaction for the hiding process so that the process needs the fewest data modifications and gives the lowest number of lost non-sensitive itemsets. The experimental results indicate that the HSUFIBL algorithm achieves better performance than previous works in minimizing the side effect.

This paper proposes a system to recognize quasi-periodic human actions from monocular video sequences. First, each input video frame is analyzed and estimated to generate the best 3D human model pose which consists of a set of 3D coordinates of specific human joints. Next, these 3D coordinates for each frame are converted into corresponding 3D geometric relational features (GRFs), which describe the geometric relations among body joints of a pose. Finally, we train a cyclic hidden Markov model (CHMM) for each action based on the vector quantized 3D GRFs, and the trained CHMMs are used to classify different quasi-periodic human actions. The experimental results indicate the effectiveness of the proposed system in terms of the view point invariance, the low-dimensional feature vectors, and the encouraging recognition rates.